NEONICOTINOIDS Country-specific effects of neonicotinoid ...€¦ · the reproductive success of bumble bees (5, 8, 10) and solitary bees ( 8, 11); other studies have iden-tified

NEONICOTINOIDS

Country-specific effects ofneonicotinoid pesticides onhoney bees and wild beesB. A. Woodcock,1* J. M. Bullock,1 R. F. Shore,2 M. S. Heard,1 M. G. Pereira,2

J. Redhead,1 L. Ridding,1 H. Dean,1 D. Sleep,2 P. Henrys,2 J. Peyton,1 S. Hulmes,1

L. Hulmes,1 M. Sárospataki,3 C. Saure,4 M. Edwards,5 E. Genersch,6

S. Knäbe,7 R. F. Pywell1

Neonicotinoid seed dressings have caused concern world-wide. We use large fieldexperiments to assess the effects of neonicotinoid-treated crops on three bee speciesacross three countries (Hungary, Germany, and the United Kingdom). Winter-sown oilseedrape was grown commercially with either seed coatings containing neonicotinoids(clothianidin or thiamethoxam) or no seed treatment (control). For honey bees, we foundboth negative (Hungary and United Kingdom) and positive (Germany) effects during cropflowering. In Hungary, negative effects on honey bees (associated with clothianidin)persisted over winter and resulted in smaller colonies in the following spring (24% declines).In wild bees (Bombus terrestris and Osmia bicornis), reproduction was negatively correlatedwith neonicotinoid residues.These findings point to neonicotinoids causing a reduced capacityof bee species to establish new populations in the year following exposure.

Global declines in honey bees and wild beeshave been linked to pathogens, climatechange, habitat fragmentation, and pes-ticide use (1–3). The potential threat fromneonicotinoid seed coatings applied to

flowering crops has been the subject of consid-erable debate (4–9). Neonicotinoids have beenshown to increase mortality in honey bees byimpairing their homing ability (4) and to reducethe reproductive success of bumble bees (5, 8, 10)and solitary bees (8, 11); other studies have iden-tified no effects (8, 12, 13). There is limited infor-mation from replicated studies on longer-termsurvival of honey bee colonies following exposure[see (12)]. Landscape-scale experiments under real-world agricultural conditions are needed to inte-grate spatial, temporal, and species-specific variationin order to understand the impacts of neonico-tinoids on bees (8, 12, 14–16). Such studies shouldexplore the impacts of different neonicotinoidformulations, land use, and regional climate. In alarge-scale experiment spanning three Europeancountries, we tested the hypotheses that (i) expo-sure to seed treatments containing neonicotinoidsaffected the reproductive potential of managedand wild bee species and (ii) whether such effectsdiffer between countries.At each of 33 sites (Germany, 9; Hungary, 12;

and United Kingdom, 12) an average of 63.1 ha(SE of ±2.8 ha) of winter-sown oilseed rape

(OSR) was established in 2014 (Fig. 1, fig. S1,and table S1). We clustered sites into triplets(>3.2 km between sites) and randomly allocatedsites to one of three treatments: (i) clothianidinapplied at 11.86 to 18.05 grams of active ingre-dient per hectare (g a.i. ha−1) with a fungicide(thriam and prochloraz) and nonsystemic pyre-throid (beta-cyfluthrin) (trade name Modesto); (ii)thiamethoxam applied at 10.07 to 11.14 g a.i. ha−1

and combined with the fungicides fludioxonil andmetalaxyl-M (trade name Cruiser); and (iii) con-trol OSR receiving a commercial fungicide (thriamand dimethomorph in Germany and Hungary andthriam and prochloraz in the United Kingdom)but no neonicotinoid seed treatment. All treatmentsreceived typical commercial inputs of pesticide(e.g., lambda-cyhalothrin) and fertilizer, with thesestandardized across a triplet. Standardized colo-nies of honey bees (Apismellifera)andwild bees (bumble beeBombusterrestris and solitary bee Osmiabicornis) were introduced to eachsite. For honey bees, we quan-tified the impacts of the treat-ments on colony viability duringthe crop flowering period and inthe year following exposure (hivesurvival and overwinteringworker,brood, and storage cell numbers).Overwintering fitness defines themultiyear persistence of honeybees. For B. terrestris, we mea-sured impacts on within-year re-productive output (colony weightgain andworker, queen, and droneproduction) and for O. bicornisthe number of reproductive cellsproduced (table S2). Neonicoti-noids can be persistent and wide-

spread in agroecosystems (17, 18), so we quantifiedresidues both in the nests of bee species andthose expressed in the crop.We found that neonicotinoid seed treatment

affected the interannual viability of honey beecolonies following the winter period in a country-specific manner. In Hungary, worker numberswere 24% lower where clothianidin was com-pared with the control [treatment × country:c2(6) = 1.47, P = 0.01, explained variance = 59.4%](Fig. 2), with no significant effect of thiame-thoxam. Clothianidin was more likely to be ex-pressed in the crop where it was applied as a seedtreatment, which identified a mechanism of expo-sure to the bees [c2(2) = 6.46, P = 0.04], but thiswas not so for thiamethoxam (table S3). In theUnited Kingdom, high hive mortality precludeda formal statistical analysis of overwinteringworker numbers. However,medianworker numberswere zero for all four clothianidin-treated sites butabove zero for two of the control and one of thethiamethoxam sites (table S2 and Fig. 2). Workernumbers following the winter in Germany showedno treatment effect (table S4). Overwintering honey-bee brood, stored hive products (pollen and nectar),and the likelihood of hives surviving the winterwere not affected by seed treatments (table S3).Neither B. terrestris queen nor O. bicornis

egg cell production was directly affected by theseed treatments or its interaction with country(table S5). However, they were negatively corre-lated with peak [c2(1) = 2.09, P = 0.03, explainedvariance = 13.5%] (Fig. 3A) and median [c2(1) =4.34, P = 0.04, explained variance = 0.8%] (Fig. 3B)neonicotinoid nest residues (combined clothianidin,thiamethoxam, and imidacloprid). Imidaclopridwas not applied as part of the study, and itspresence is most likely a result of environmentalcontamination from previous widespread agro-nomic use (17, 18). Residues of neonicotinoidsdetected in stored hive products did not differ inresponse to seed treatments for any bee species(table S6). This may be due to the amalgamationof stored hive products at the site level for resi-due analysis, which may have obscured within-siteheterogeneity in residues. The negative correlation

RESEARCH

Woodcock et al., Science 356, 1393–1395 (2017) 30 June 2017 1 of 3

1Centre for Ecology and Hydrology, Natural EnvironmentResearch Council, Oxfordshire OX10 8BB, UK. 2Centre forEcology and Hydrology, Natural Environment Research Council,Lancaster Environment Centre, Lancaster LA1 4AP, UK. 3SzentIstván University, 2103 Gödöllö, Hungary. 4Am-Heidehof 44,14163 Berlin, Germany. 5Leaside, Carron Lane, West SussexGU29 9LB, UK. 6Institute for Bee Research, 16540 Hohen-Neuendorf, Germany. 7Eurofins, Ecotox-GmbH, 75223 Niefern-Öoschelbronn, Germany.*Corresponding author. Email: [email protected]

Fig. 1. Location of the 33 experimental sites in the UnitedKingdom, Hungary, and Germany. See fig. S2 for a diagrammaticrepresentation of the experimental setup.

on June 29, 2017

http://science.sciencemag.org/

Dow

nloaded from


for B. terrestris queen production remained sig-nificant when we excluded sites with imidaclopridresidues [c2(1) = 2.14, P = 0.02], although this wasnot the case for O. bicornis [c2(1) = 0.05, P = 0.81].Country-specific responses to neonicotinoid seedtreatment were found for B. terrestris drone pro-duction, with positive and negative effects fromexposure to thiamethoxam in Germany and theUnited Kingdom, respectively [treatment × country:c2(6) = 13.1, P = 0.04, explained variance = 13.6%](Fig. 2).We also found seed treatment effects during the

crop flowering period that lasted between 3 and6 weeks (tables S4 and S5). Significant interactionsbetween seed treatment and country were iden-tified for peak worker [c2(6) = 16.6, P < 0.01, ex-plained variance = 45.3%], egg cell [c2(6) = 4.13,P = 0.01, explained variance = 49.9%], and com-bined pollen and nectar storage cell [c2(6) = 40.5,P < 0.001, explained variance = 53.6%] numbers.These responses describe within-year colony per-formance. Neonicotinoid exposure resulted inboth negative (Hungary and United Kingdom)and positive (Germany) effects on colony size(see Fig. 2; pairwise treatment comparisongiven in tables S4 and S5). Bombus terrestrisworker and peak colony weight showed noseed treatment response.Our quantification of neonicotinoid effects on

the interannual viability of honey bees and wildbee populations represents a fundamental advancein our understanding of the impacts of thesepesticides. For solitary bees and bumble bees(queen production), neonicotinoid impacts wereassociated with the residues found in nestsrather than the experimental seed treatments.For B. terrestris, the few treatment effects andthe presence of imidacloprid in stored pollenand nectar (tables S7 to S9) suggests that neg-ative impacts of neonicotinoids may be drivenby the persistence of residues in the wider land-scape rather than current management alone(18, 19). The European Union (EU) moratoriummeant that no neonicotinoids were applied tooilseed in the surrounding landscapes duringthe experiment, so such residues may originatefrom previous agricultural use leading to ex-pression in nontarget plants (17–19), guttationfluids, or contaminated water (19, 20). Althoughthe reproductive potential of O. bicornis wasalso negatively affected by neonicotinoid resi-dues in nests, the explained variation of theseeffects was small. However, a failure to detectsmall population changes may be due to limitedexperimental replication restricting statisticalpower. Our results suggest that even if theiruse were to be restricted, as in the recent EUmoratorium, continued exposure to neonico-tinoid residues resulting from their previouswidespread use has the potential to impact neg-atively wild bee persistence in agricultural land-scapes (14, 18, 19).Taken together, our results suggest that expo-

sure to neonicotinoid seed treatments can havenegative effects on the interannual reproductivepotential of both wild and managed bees, butthat these effects are not consistent across coun-

tries. The country-specific responses of honeybees and bumble bees strongly suggest that theeffects of neonicotinoids are a product of inter-acting factors (20–23). This study has identifiedbetween-country differences in the use of oilseedrape crop as a forage resource for bees (affect-ing exposure to crop residues) and incidenceof disease within hives. Both factors were high-er for Hungarian and U.K. honey bees (tablesS10 and S11). Overall neonicotinoid residues

were detected infrequently and rarely exceeded1.5 ng g−1 (w/w). As such, direct mortality effectscaused by exposure to high concentrations ofneonicotinoids are likely to be rare (table S12).However, our results suggest that exposure tolow levels of neonicotinoids may cause reductionsin hive fitness that are influenced by a number ofinteracting environmental factors. Such interact-ing environmental factors can amplify the impactof honey bee worker losses (e.g., through sublethal


Fig. 2. Summary effect sizes for the response of honey bees and wild bees to the neonicotinoidseed treatments. An effect size represents the difference between the mean population response for agiven seed treatment and the control within a country, with this difference divided by the pooledstandard deviation, where asterisk (*) indicates a significant difference between the control and seedtreatment [either TMX (thiamethoxam) or CTD (clothianidin)] determined from the predicted marginalmeans of the model “y ~ seed treatment × country + block/country.” Dagger (†) indicates where U.K.colony survival was too low for a formal analysis. Note that effect sizes differ between countries.

Fig. 3. Wild bee reproductive success in response to neonicotinoid nest residues. Separategraphs are shown for the response of B. terrestris queen production and O. bicornis reproductive cellproduction to neonicotinoid residues found in nests. The significance of these relationships is basedon a likelihood ratio test comparison of H0: “y ~ country” and H1: “y ~ neonicotinoid + country.”Neonicotinoid residues are based on summed concentrations of clothianidin, thiamethoxam, andimidacloprid.

RESEARCH | REPORTon June 29, 2017


Dow

nloaded from


toxicity effects) and reduce longer-term colonyviability (4, 16). Note that our common experimen-tal approach applied across three countries revealedvarying impacts and may explain the inconsistentresults of previous studies conducted in singlecountries or at few sites (4, 5, 8, 12, 13, 15).

REFERENCES AND NOTES

1. A. J. Vanbergen, Front. Ecol. Environ 11, 251–259 (2013).2. S. G. Potts et al., Trends Ecol. Evol. 25, 345–353 (2010).3. R. Winfree, R. Aguilar, D. P. Vázquez, G. LeBuhn, M. A. Aizen,

Ecology 90, 2068–2076 (2009).4. M. Henry et al., Science 336, 348–350 (2012).5. P. R. Whitehorn, S. O’Connor, F. L. Wäckers, D. Goulson,

Science 336, 351–352 (2012).6. J. E. Cresswell et al., Zoology 115, 365–371 (2012).7. B. A. Woodcock et al., Nat. Commun. 7, 12459 (2016).8. M. Rundlöf et al., Nature 521, 77–80 (2015).9. G. E. Budge et al., Sci. Rep. 5, 12574 (2015).

10. D. Goulson, PeerJ 3, e854 (2015).11. C. Sandrock et al., Agric. For. Entomol. 16, 119–128 (2014).12. G. C. Cutler, C. D. Scott-Dupree, M. Sultan, A. D. McFarlane,

L. Brewer, PeerJ 2, e652 (2014).13. G. Christopher Cutler, C. D. Scott-Dupree, Ecotoxicology 23,

1755–1763 (2014).14. B. A. Woodcock et al., J. Appl. Ecol. 53, 1358–1362 (2016).15. E. Pilling, P. Campbell, M. Coulson, N. Ruddle, I. Tornier,

PLOS ONE 8, e77193 (2013).16. M. Henry et al., Proc. R. Soc. B Biol. Sci. 282, 2015.2110 (2015).17. A. Jones, P. Harrington, G. Turnbull, Pest Manag. Sci. 70,

1780–1784 (2014).18. C. Botías et al., Environ. Sci. Technol. 49, 12731–12740 (2015).19. D. Goulson, J. Appl. Ecol. 50, 977–987 (2013).20. A. Fairbrother, J. Purdy, T. Anderson, R. Fell, Environ. Toxicol.

Chem. 33, 719–731 (2014).21. FERA, Neonicotinoid Pesticides and Bees. Report to Syngenta

Ltd. (The Food and Environment Research Agency, UK, 2013).22. F. Sánchez-Bayo et al., Environ. Int. 89-90, 7–11 (2016).23. C. R. Archer, C. W. W. Pirk, G. A. Wright, S. W. Nicolson, Funct.

Ecol. 28, 913–923 (2014).

ACKNOWLEDGMENTS

Data are in supplementary materials. Funded by Syngenta Ltd.and Bayer CropScience (P. Campbell, M. Miles, C. Maus, D. Holah,M. Coulson). Wild pollinator work supported by NERC CEHNational Capability funding (NEC05829). Thanks to K. Jaekel,P. Fisher, M. Nowakowski, R. Hails, P. Scrimshaw, N. Mitschunas,P. Nuttall, M. McCracken, S. Ball, J. Webb, B. Sutherland,R. Freckleton, T. Tscharntke, J. Memmott, K. Norris, B. Raffa,and D. Vaskor.

SUPPLEMENTARY MATERIALS

www.sciencemag.org/content/356/6345/1393/suppl/DC1Materials and MethodsFigs. S1 and S2Tables S1 to 12References (24–32)

7 December 2016; accepted 22 May 201710.1126/science.aaa1190


RESEARCH | REPORTon June 29, 2017


Dow

nloaded from

http://www.sciencemag.org/content/356/6345/1393/suppl/DC1


www.sciencemag.org/content/356/6345/1393/suppl/DC1

Supplementary Materials for

Country-specific effects of neonicotinoid pesticides on honey bees and wild bees

B. A.Woodcock,* J. M. Bullock, R. F. Shore, M. S. Heard, M. G. Pereira, J. Redhead,

L. Ridding, H. Dean, D. Sleep, P. Henrys, J. Peyton, S. Hulmes, L. Hulmes, M. Sárospataki, C. Saure, M. Edwards, E. Genersch, S. Knäbe, R. F. Pywell

*Corresponding author. Email: [email protected]

Published 30 June 2017, Science 356, 1393 (2017) DOI: 10.1126/science.aaa1190

This PDF file includes

Materials and Methods Figs. S1 and S2 Tables S1 to S12 References

Materials and Methods In August-September 2014, large continuous blocks of winter sown oilseed rape

(mean=63.1 ha, SE=2.8; range 31.3 – 95.6 ha) were established on 33 commercial sites split between the UK (12 sites), Hungary (12 sites) and Germany (9 sites) (Table S1). These countries represent a gradient in climate and environmental characteristics resulting from historical land (e.g. collectivization, field size and landscape diversity) and agri-management practices (pesticide use and cropping rotations). By using a common experimental framework this study will test whether the impact of neonicotinoid seed treatments is consistent among the wide range of environmental conditions represented by the three studied countries. Winter sown oilseed rape was chosen as it represents a key element of the cropping system in Europe (24). The far less widely sown spring oilseed rape (as tested in Rundlöf et al (8)) has higher expression of neonicotinoids (> 10 ng g–1) than winter sown varieties (> 3.5 ng g–1 (13)). The area of sown oilseed rape at each site reflects typical block cropping patterns of commercial farming systems. In each country the variety of oilseed rape was chosen from popular hybrid varieties (UK – var. Harper; Germany – var. Flyer; Hungary – var. Hybrirock). Sites were clustered into blocks of three sites located in similar landscapes and on similar soils within each country. Sites were separated on average by 5.47 km (SE=0.54, minimum = 3.2 km), with blocks separated by > 10 km.

Within each block (containing three sites, see Fig. S1) one site was randomly allocated to each of the following oilseed rape neonicotinoid seed treatments: 1) Clothianidin (Modesto®, Bayer Cropscience Ltd.) achieving a within field application rate of 11.86 g ha–1 a.i. (UK), 18.05 g ha–1 a.i. (Germany) and 17.71 g ha–1 a.i. (Hungary). Modesto® is combined with a fungicide (Thriam and prochloraz) and non-systemic pyrethroid (beta-cyfluthrin) used to protect seeds against immediate pest damage on sowing; 2) Thiamethoxam (Cruiser®, Syngenta Ltd.) achieving an application rate of 10.07 g ha–1 a.i. (UK), 10.61 g ha–1 a.i. (Germany) and 11.14 g ha–1 a.i. (Hungary). Cruiser® is combined with a fungicide (fludioxonil and metalaxyl-M; 3) Control oilseed rape where seeds received only a commercial fungicide (Thriam and Dimethomorph (Germany & Hungary) or Thriam and prochloraz (UK)), but no neonicotinoid seed treatment. All treatments received typical commercial inputs of pesticide (e.g. Lambda-cyhalothrin or Tau-fluvalinate) and fertiliser which were standardized within a block. Sites were chosen so that there was no other oilseed rape crop except for that established as part of the experiment within 1.5 km of the hives. Due to the EU moratorium any oilseed rape outside of this radius would not have been treated with neonicotinoids. Honey bees

The domesticated honey bee represent one of the principal pollinators of oilseed rape (4, 12, 15). At each site six honey bee hives were located centrally to the oilseed rape crop. In the majority of cases where multiple fields of oilseed rape were sown, hives were placed as close to the field centers as possible along an existing field boundary. Where a single large field of oilseed rape was sown, a 50 x 100 m area was cleared in the center of the crop and the hives were placed there. Hives were established at the BBCH 61 developmental stage for each block (5-11/5/2015 in Germany, 14-21/4/2015 in Hungary and 9-23/4/2015 in the UK). Within a country honey bees originated from the

2

same genetic stock. For Germany and Hungary hives were 1 year old colonies of the carnica sub species (average pre-exposure colony size: Germany=10683 workers, SE=405; Hungary= 8993, SE=248). In the UK it was not possible to source sufficient full size colonies with the same provenance and so new (6 frame) nuclei colonies were produced from young queens (1 year old) of the ‘Buckfast’ strain (average of 3294 workers, SE±126). Following the end of the oilseed rape flowering period, hives were removed to a single overwintering site within each country in a landscape that provided non-crop flowering resources (Germany: Lat. 50.763802, Long. 9.099248; Hungary: Lat. 47.335223, Long. 20.311399; UK: Lat. 51.888052, long. -1.4344352). As all honey bees within a country were exposed to a common set of conditions at these sites (e.g. foraging resources, landscape and weather) their population sizes as monitored the following year were the result of the experimental treatments during the previous oilseed rape flowering period. Bees were supplementary fed with fondant or sugar solution depending on typical practice in each country. Hives were treated for Varroa using either formic acid (Germany and Hungary) or oxalic acid (UK). However, country specific differences in disease rates were observed, with the UK having higher levels of Varroa mite infection (8.05 % of worker bees ±SE 1.34) than those of either Germany (1.04 % ±SE 1.00) or Hungary (2.12 % ±SE 1.34) (χ2

2=16.6, p<0.001; Table S11). In the UK, honey bees had a narrower diet breadth (fewer plant species present in pollen samples) than those of either Germany or Hungary (χ2

2=9.81, p= 0.001; Table S10). Honey bees in the UK and Hungary also foraged on higher proportions of pollen originating from oilseed rape than Germany (χ2

2=12.2, p<0.001; Table S10). The impact of honey bee exposure to neonicotinoid seed treatments was assessed

using standard Liebefeld (25) counts of worker, egg cell, larvae, pupae, male brood and combined storage cells (pollen and nectar) within each hive. Male brood numbers were too infrequent for a valid statistical analysis. Assessments were undertaken during the oilseed rape flowering period (April-June 2015) and again during the post-winter period (March 2016). In the oilseed rape flowering period assessments were made at 4-7 days after deployment at sites and then again at weekly intervals until the end of flowering. The flowering period differed between countries lasting from 3 (Hungary) to 6 weeks (Germany and UK). Over this period the peak numbers of bees and brood developmental stages were assessed. However, to ensure that peak counts reflected responses to the oilseed rape crop the first sampling round (undertaken at 4-7 days) was ignored. Hive weight was adjusted so that the first colony weight measured during the flowering period was set to zero. The final colony assessment in the oilseed rape flowering period was undertaken on 21/5/2015 in the UK, 12/5/2016 in Hungary and 8/6/2016 in Germany. Overwintering survival and colony strength was also assessed using the Liebefeld method. This was undertaken on single date in March 2016 (UK: 30/3/2016; Hungary: 8/3/2016; Germany: 15/3/2016). Following the overwintering period individual life stages of egg, larvae, pupae and male brood were infrequent and so were summed to provide a measure of total brood. In all cases median site values from the six hives at a site were used in subsequent analyses.

During the oilseed rape exposure period samples of pollen and nectar collected by the honey bees were analysed for residues of clothianidin, thiamethoxam and imidacloprid (a third widely used neonicotinoid product). While this latter neonicotinoid was not applied as part of the study, its residues have been reported both in arable soils

3

(17, 18). At each site residues of these three products were determined from: 1) pollen stored in combs within the hives (sampled from 10 cells from each of six hives at a site and then homogenized); 2) nectar stored in combs (sampled as for stored pollen); 3) nectar collected by honey bees and dissected from their honey stomachs on two occasions during the flowering period (300 bees from six hives at a site); 4) Pollen collected by worker bees using pollen traps and sampled on three occasions (taken from six hives at a site and homogenized). Prior to overwintering (November 2015) pollen and nectar stored in combs within the hives were again analysed for residues using the same method. These approaches provide information on residues from a range of sources of pollen and nectar at a site level, but do not attempt to quantify within hive spatial variability of neonicotinoid residues in stored hive products. By amalgamating samples originating from multiple hives within a site it is possible that detection rates of neonicotinoids may be reduced where high levels of within and between hive spatial variability in the distribution of residues exists. Bumblebees (Bombus terrestris)

This bumblebee regularly visits flowers of oilseed rape and is an important crop pollinator worldwide (26). At each study site 12 commercial B. terrestris colonies (Biobest Ltd.) of either audax (UK) or terrestris sub-species (Hungary and Germany) were established reflecting distributional ranges and regulatory issues. The 12 colonies were clustered into four multi-hives (a multihive is a cluster of three colonies housed in the same protective polystyrene box) and were placed central to the sown oilseed rape crop. As for the honey bees, this was either along a field boundary where multiple fields were sown with oilseed rape, or in a cleared area in the center of crop with only a single large field sown per site. The entrance to individual colonies was restricted to prevent parasitism by B. psythirus s.g.. This also prevented the escape of new queens produced by the colonies. Bombus colonies were places in fields at the same time as honey bee hives. The pre-exposure colony size was a mean of 102.2 workers (SE=1.9) in Germany, 81.2 (SE=1.06) in Hungary and 93.6 (SE=1.36) in the UK.

Bumblebee colonies do not produce reproductive stages (queens and drones) until after the end of oilseed rape flowering. At this stage they have typically consumed all stored hive products. For this reason colonies were collected in two phases. The first 6 colonies (2 multihives) were collected at the end of the oilseed rape flowering period (UK: 20/5/2015; Hungary: 18-19/5/2016; Germany: 30/5/2015 – 1/6/2016) in order to measure neonicotinoid residues in stored hive products (pollen and nectar). In addition, pollen was collected from the pollen baskets of workers returning to multihives.

The remaining six colonies were collected after 51-60 days following their exposure to the treated crop (UK: 9-11/6/2015; Hungary: 17-18/6/2016; Germany: 20-21/6/2016) in order to measure effects on reproductive success. Each colony was dissected and the total number of workers, queens and drones were counted. Queen number was defined as the combined number of emerged queens and un-emerged queen cells which could be distinguished on the basis of their larger size (>12 mm) (8). As the aperture entrance to colonies was too small to allow newly emerged queens from leaving counts of queens are likely to be very accurate. It was not possible to prevent drones from leaving the hives as they are similar in size to workers. As larval worker and drone cells were equivalent in size it was not possible to incorporate pupal cell numbers in the

4

estimation of either of these two castes. As there was typically considerable spill-over between colonies within a multihive counts of all castes were summed within a single multihive. Median numbers of developmental stages (queens, workers and drones) per site were used as a response. In addition to the dissections of hives, individual multihives were weighted at the same time at the honey bees at weekly intervals during the oilseed rape flowering period.

Solitary bees (Osmia bicornis)

This solitary bee has been identified as an important contributor to the worldwide value of crop production and frequently feeds on oilseed rape (11, 26). At each site 50 Osmia bicornis cocoons containing an equal ratio of males to females were placed in protected release cage. These were located next to two artificial trap nests (standardized wooden blocks (11 × 11 × 28 cm) drilled with equal number of 4, 6, 8 and 10mm diameter holes) attached to 2 m poles to provide nesting locations. Osmia species naturally nest along field edges so cocoons and trap nests were always situated at the edge of each fields. At the end of the oilseed rape flowering period (June 2015) the two trap nests were dissected and counts of the number of O. bicornis cells were made. A median count of the number of cells per site was used in subsequent analyses. Pollen and nectar samples were removed from individual cells to test for residues of neonicotinoids. Note, that as no O. bicornis reproductive cells were produced at three sites no pollen and nectar could be collected for analysis of neonicotinoid residues. Neonicotinoid expression in the oilseed rape crop

To quantify the expression of neonicotinoids found within the treated and untreated crops a single honey bee hive was placed within a fine mesh cage (20m long × 5m wide × 3m tall) over the flowering crop. Honey bees were thus used as vectors to collect pollen and nectar from a large number of flowers at each site. This method has the advantage that it simultaneously collects pollen and nectar samples from a large number of oilseed rape plants. However, should there be spatial variation in the expression of neonicotinoids between plants some of this variation may be averaged out as the pollen and nectar collected from all plants were amalgamated. Pollen was collected from honey bees using pollen traps placed on the front of the hive. To collect nectar 300 bees returning to hives were collected and nectar was removed from their honey stomachs by dissection. Nectar and pollen was collected on two occasions (day 2 and day 9 of crop flowering). There was evidence of unexpected expression of neonicotinoids in the crop that did not correspond with the seed treatments (Table S12). Imidacloprid was detected in three sites. Clothianidin residues were found in one control and two thiamethoxam sites, although as clothianidin is a metabolite of thiamethoxam this does not reflect contamination (21). Thiamethoxam was found in one of the control sites and a clothianidin treated site. Although infrequent, unexpected expression in the crop was linked to historic patterns of pesticide application on these sites (17). Neonicotinoid residue analysis

All pollen and nectar samples were transported in insulated boxes with dry ice and stored long term at -80°C. Samples of pollen and nectar were analysed to quantify concentrations of clothianidin, thiamethoxam (UKAS accredited ISO17025:2005

5

standards) and imidacloprid. Analysis was performed using a liquid chromatography coupled to a triple quadrupole ‘Quantum Ultra TSQ’ mass spectrometer (Thermo Fisher Scientific, Hemel Hemsptead; UK) interfaced with an ion max electrospray ionisation (ESI) system (see http://www.ceh.ac.uk/sites/default/files/NNI%20Residue%20method%20technical%20protocol_0.pdf for full protocol). A Limit of Quantification (LoQ) for both pollen and nectar samples of 0.53 ng g–1 (Limit of Detection (LoD) = 0.38 ng g–1) was obtained for samples from the honey bee and B. terrestris. For O. bicornis the LoQ was 0.52 ng g–1 (LoD = 0.37 ng g–1). Residues below the LoQ were defined in the data set to be half LoD. The LOD was determined using 3 times the signal to noise ratio while the LOQ was calculated as the LOD plus the calculated expanded uncertainty using the method defined by Magnusson et al (27).

Median and maximum residues of clothianidin, thiamethoxam and imidacloprid were determined for each site based on the multiple pollen and nectar samples. As neonicotinoid residues were detected in sites where that product was not applied a combined metric of clothianidin, thiamethoxam and imidacloprid was derived (NNImedian and NNIMax). While all three neonicotinoids have a similar mode of action (21, 28) they do differ slightly in toxicity (29, 30, 31). To account for this we provide a standard correction for the summed concentration of neonicotinoids based on acute oral LD50 values for honey bees (TMX=0.005 μg bee–1; CTD= 0.00379 μg bee–1; IMI=0.0037 μg bee–1 , or a ratio of their relative oral toxicity of 1: 0.758: 0.74) (29, 30, 31) so that NNI= (TMX) + (CPD × 0.758) + (IMI × 0.74). This same correction was used for wild pollinators as comparable LD50 values were not available for these species.

Landscape structure

Landscape composition was recorded within a 1.5-km radius of the center of each site based on direct field observations of habitat types derived from onsite surveys undertaken during the oilseed rape flowering period. Each discrete land parcel (minimum size 25 × 25 m) was digitized using ARCGIS software and assigned to: crops (oilseed rape, cereals, legumes or vegetables), horticulture, grassland (improved, semi-Improved, unimproved), woodland (linear feature, broadleaf, coniferous or mixed), scrub, water or urban land use. We include as covariates in subsequent model percentage area of oilseed rape (to describe the potential resource on which bees could forage) and the percentage cover of all arable crops. The cover of arable crops acts as a measure of agricultural land use intensity while the cover of oilseed rape defines the availability of the crop foraging resource.

Statistical Analysis

The analysis aimed to test whether population metrics for honey bees, B. terrestris and O. bicornis differed in size relative to the control where clothianidin or thiamethoxam was used to treat oilseed rape. All analyses were undertaken using gerneralized linear models in R 3.3.1 (32). These models are specified in full in Tables S4 and S5. The analysis was a two-step process. Firstly we tested whether continuous covariates describing between site variations in environmental conditions (landscape structure) and neonicotinoid exposure risk explained additional variation over that seen

6

http://www.ceh.ac.uk/sites/default/files/NNI%20Residue%20method%20technical%20protocol_0.pdf

http://www.ceh.ac.uk/sites/default/files/NNI%20Residue%20method%20technical%20protocol_0.pdf

for a country only model. This was done separately for covariate describing neonicotinoid residues in the nests (natural logs of NNImedian and NNIMax), neonicotinoid residues expressed in the oilseed rape crop (natural logs of NNIMax) and landscape percentage cover of oilseed rape and arable crops. The significance of individual covariates was assessed using a likelihood ratio test comparing each covariate (y ~ covariate + country) to a simple country only model (y ~ country). Other potential covariates showed strong systematic variation with either country or seed treatment and as such could not be directly tested within this framework without violated underlying model assumptions (e.g. the proportion of oilseed rape making up the diet of pollen (Table S10), disease (Table S11) and starting colony size (Table S2)).

Following this, we then tested whether these same metrics showed either no response to the seed treatments, an overall seed treatment response common to all countries, or a seed treatment response that was country-specific. This was done by undertaking three likelihood ratio tests (LRT) for each population metric. These were: (i) a test to assess whether a simple overall seed treatment effect common to all countries explained more variance than a country only model (LRT comparing ‘y ~ treatment + covariates + block/country’ to the null model ‘y ~ covariates + block/country’); (ii) a test to assess if there was an additive seed treatment effect that was unique to each country (LRT comparing ‘y ~ treatment*country + covariates +block/country’ to the null model ‘y ~ covariate + block/country’); and (iii) a test to assess if the additive seed treatment effect describing country specific responses was a better fit to the data than a simple overall seed treatment effect (LRT comparing ‘y ~ treatment*country + covariates + block/country’ to the simple seed treatment only model ‘y ~ treatment + covariates + block/country’). Covariates were included only where they were demonstrated to explain additional variance in the population metrics over a simple country only model as explained above. It should be noted that none of the significant responses to seed treatment were not dependent on the inclusion of these continuous covariates within the models. Count data were typically over-dispersed and were modelled using a negative binomial distribution and log link (glm.nb function in the MASS package in R). Where response variables were normally distributed (assessed using a Shapiro-Wilk normality test confirmed by QQ plots of theoretical and sample quantiles) a normal distribution with identify link was used. Honey bee colony survival was modelled using an events / trials approach with binomial errors and logit link. Individual model-predicted marginal means were used to assess the significance of differences between the control and seed treatments within country (Tables S4 and S5). To confirm models met underlying assumptions visual checks of standard residual diagnostic plots were undertaken. We did not apply Bonferroni corrections as the lack of independence for the majority of the response variables (e.g. different life stages of honey bee) meant that there was no valid level for the correction. Note that honey bee overwintering survival in the UK was considered too low for valid statistical testing. As such the analysis of overwintering honey bee colony strength was restricted to Germany and Hungary data sets only.

A further analysis was undertaken to assess if the occurrence of neonicotinoid

residues in stored hive products (pollen and nectar) varied in response to seed treatment. This analysis followed the same structure described above with full models specified in Table S6. Individual clothianidin, thiamethoxam and imidacloprid residues were too

7

infrequent to provide a robust analysis of their individual responses (Table S7 to S9). We therefore assessed the response of the combined index (NNIMax) of neonicotinoid residues to the additive seed treatment and country specific seed treatment effects models as described above. In addition to this, we also tested using the same approach whether the expression of neonicotinoid pesticides within the oilseed rape crop could be predicted by seed treatment (see Table S3 for individual models). Site median and peak residues of clothianidin and thiamethoxam (based on the four combined samples per site of pollen and nectar) had a large number of non-detect values making a direct analysis of these not possible (Table S12). To account for this we undertook an events (number of pollen and nectar samples with residues > LoQ): trials (total number of pollen and nectar samples) analysis using binomial errors (logit link). This assessed the proportion of pollen and nectar samples from the crop that had residues above the LoQ of 0.53 ng g–1 w/w. We note that while this approach is suitable for identifying broad patterns in the expression of neonicotinoids in the crop it does not take into account different levels of expression of these compounds in pollen and nectar. However, given the paucity of residue data the same analyses treating pollen and nectar residues separately would be too data poor to allow robust inferences (individual pollen and nectar residues are given in Tables S7, S8, S9 and S12). Power analysis

Figures S2A-C presents a power analysis used to assess the capacity of the experimental design to detect overall additive neonicotinoid seed treatment effects that were consistent across all countries (i.e. y ~ seed treatment + block/country). Figure A (honey bee response metrics recorded during the oilseed rape flowering period), B (honey bee population metrics following the overwinter period) and C (wild bee population metrics) provides the power (1-β) of the experimental design to detect effect size reductions from 7 – 50 % under different levels of replication (3, 4, 5, 10 or 20 replicate blocks within each country). To perform the power analysis we simulated data to determine how the number of replicate blocks within each country (where a block represents a control, clothianidin and thiamethoxam treated site) affects the power (1-β) of the study to detect a given effect size difference between the control and a single neonicotinoid seed treatment. This reflects the key experimental goal of identifying consistent cross-country effects of neonicotinoid seed treatments. To perform the power analysis we used a simple country only null model to derived estimates for univariate distributions, e.g. mean and standard deviation for normal distribution or theta for Negative binomial. From this we simulated data sets where effect size difference was altered between the control and a single neonicotinoid seed treatment. This was done by generating random outcomes from either a negative binomial, normal or binomial distribution depending on the response variable (Table S4 and S5). For each combination of effect size and number of blocks 10,000 simulations were run. Power (1-β) was defined to be the proportion of these simulations (for each effect size and block number combination) that had a probability of detecting an additive seed treatment effect.

8

Fig S1.

Osmia trap nests

Treatment Clothianidin

Treatment Thiamethoxam

> 10 km separation

2) Replicate blocks within country

UK Germany Hungary

1) Country scale replication

> 3.2 km separation

Control No

neonicotinoids

3) Within block treatments applied at random

4) Within site layout

Sown OSR c.60 ha per site (Either into single fields or

group of fields)

Honey bee

Tents used to assess crop residue expression

Bombus terrestris

OS

OS

OS

OS

Surrounded by crops not attractive to bees to at least 1.5km from hives

9

A simple diagrammatic representation of the experimental set-up. This shows the three seed treatments applied to winter oilseed rape (clothianidin, thiamethoxam and control). Each seed treatment was sown at the scale of an individual site (e.g. a farm), which represents our experimental unit. Triplets of sites were nested within blocks, with these blocks replicated within each of the three countries (UK and Hungary = 4 blocks, Germany = 3). Within each site honey bee hives, B. terrestris colonies and O. bicornis populations were established as close to the center of the sown oilseed rape as possible (although the three species were always separated by 50 m to limit direct interference). For some sites multiple fields of oilseed rape were needed to achieve the target coverage of crop. At these sites all three species were placed on a field boundary close to the centroid of the sown crop. At some sites only a single large field was sown. At these sites an area in the center of the field was cleared of the crop and the honey bee and B. terrestris colonies were placed here. For O. bicornis trap nests were always located on boundaries as the behavior of these species would prevent them colonizing trap nests in the center of a large field. The cages used to assess the expression of neonicotinoid residues in the crop were also established on one of the sown oilseed rape fields at each site.

10

Fig. S2. (A)

(a) Peak worker (b) Peak egg (c) Peak larvae

Pow

er (1

-β)

(d) Peak pupae (e) Peak storage cell

Number of blocks per country

Detectable effect size (%)

11

(B)

(a) Overwintering worker number

(b) Overwintering brood cell number

(c) Overwintering storage cell number

Pow

er (1

-β)

(d) Colony overwintering survival


Detectable effect size (%)

12

(C)

(a) B. terrestris colony queen production

(b) B. terrestris colony drone production

(c) B. terrestris colony worker number

Pow

er (1

-β)

(d) B. terrestris peak colony weight

(e) O. bicornis cell production.


Detectable effect size (%) Fig. S2. Summary power analysis for individual population metrics. Summary graphs from a power analysis used to assess the capacity of the experimental design to detect overall additive neonicotinoid seed treatment effects that were consistent across all countries (i.e. y ~ seed treatment + block/country). (A) Honey bee response metrics recorded during the oilseed rape flowering period. (B) Honey bee population metrics following the overwinter period. (C) Wild bee population metrics. These data provide the power (1-β) of the experimental design to detect effect size reductions from 7 – 50 % under different levels of replication (3, 4, 5, 10 or 20 replicate blocks within each country).

13

Table S1. Site characteristics. A description of individual site characteristics including allocation of sites to neonicotinoid seed treatments (where CTD = clothianidin, TMX= thiamethoxam) and separation from the next closest site. Cover of oilseed rape and all arable crops within 1.5-km radii of hives is provided.

Country Site / block

Treat. Nearest site (km)

Sown area (ha) (Percentage cover

OSR)

Cover of arable

crops (%)

Shannon habitat

diversity

Germany 1 (Bl.1) Control 9.3 66.7 (9.7) 34.4 1.37

Germany 2 (Bl.1) CTD 4.1 95.7 (17.2) 33.6 1.67

Germany 3 (Bl.1) TMX 4.1 74.3 (15.3) 65.3 1.44


Germany 5 (Bl.2) CTD 7.0 37.4 (5.7) 50.8 1.66

Germany 6 (Bl.2) TMX 7.0 31.4 (10.4) 74.7 1.20


Germany 8 (Bl.3) CTD 5.9 54.6 (12.9) 93.6 0.73

Germany 9 (Bl.3) TMX 4.8 74.3 (16.0) 46.6 1.57

Hungary 10 (Bl.1) Control 3.3 47.5 (7.0) 93.8 1.32

Hungary 11 (Bl.1) TMX 3.3 55.7 (12.1) 90.7 1.47

Hungary 12 (Bl.1) CTD 11.6 43.7 (17) 87.5 1.53


Hungary 14 (Bl.2) TMX 4.5 46.1 (15.3) 66.5 1.68

Hungary 15 (Bl.2) CTD 4.5 50.6 (9.8) 42.5 1.65


Hungary 17 (Bl.3) TMX 3.7 77.4 (11.1) 65.5 1.63

Hungary 18 (Bl.3) CTD 3.7 76.7 (14.5) 58 1.28

Hungary 19 (Bl.4) CTD 10.8 64.8 (12.6) 98.7 1.4


Hungary 21 (Bl.4) TMX 4.7 53.6 (9.3) 61.3 1.59

UK 22 (Bl.1) TMX 4.6 85.8 (12.6) 35.7 1.72

UK 23 (Bl.1) Control 4.6 75.1 (14.1) 35.5 1.75

UK 24 (Bl.1) CTD 6 56.7 (9.4) 75.9 1.9

UK 25 (Bl.2) Control 8.2 56.9 (33.2) 72.6 1.8

UK 26 (Bl.2) TMX 10.5 70.6 (8.7) 44.8 2.06

UK 27 (Bl.2) CTD 8.2 71.2 (15.9) 63.5 1.79

UK 28 (Bl.3) CTD 12 91.0 (10.2) 45.8 1.98

UK 29 (Bl.3) Control 12 56.7 (12.8) 61.6 1.77

UK 30 (Bl.3) TMX 17.3 73.1 (8.0) 51.3 1.81

UK 31 (Bl.4) Control 4.7 59.6 (10.7) 25.3 1.88

UK 32 (Bl.4) TMX 7.5 61.4 (8.6) 37.1 1.79

UK 33 (Bl.4) CTD 4.7 36.9 (9.0) 47.6 2.31

14

Table S2. Peak honey bee population metrics derived during the flowering (S2A) and overwintering (S2B) periods. Reproductive success of B. terrestris and O. bicornis (S2C). Presented values are based on site medians across hives and colonies for honey bees and B. terrestris. For this reason the number of hives (out of six) surviving the winter period can be above zero while median population sizes for that site can be zero. Peak hive / colony weights for honey bees are given relative to the first weight measurement when hives were placed out on experimental sites. Table S2A

Site (treatment)

Workers (pre-exposure to

crop)

Peak worker

Peak egg

Peak larvae

Peak pupae

Peak male

brood

Peak storage

cell

G1 (Cont.) 11737 21850 4960 7920 23040 1120 122480

G2 (CTD) 10050 21975 6080 8400 23680 1120 88480

G3 (TMX) 12150 22350 4080 5920 23040 480 125760

G4 (Cont.) 12375 16550 2960 2560 15040 240 108160

G5 (CTD) 13350 23025 5440 8320 20080 880 119040

G6 (TMX) 7987 22000 9120 8800 23040 640 79600

G7 (Cont.) 9975 22250 3280 6720 23680 1680 88160

G8 (CTD) 9487 16325 4880 7200 24080 1280 89360

G9 (TMX) 7275 19800 5360 8080 24240 1200 86960

H10 (Cont.) 8712 19380 3672 9112 30328 0 37672

H11 (TMX) 7905 19508 2856 8296 25160 0 53040

H12 (CTD) 7990 17723 3264 11560 24480 0 52904

H13 (Cont.) 8585 23037 6800 10472 30192 0 51620

H14 (TMX) 11517 27507 5576 10336 31824 136 57528

H15 (CTD) 7947 20995 5712 10744 28696 0 50592

H16 (Cont.) 8925 25683 6120 12104 34272 0 50320

H17 (TMX) 10285 29745 4216 8568 30736 0 53176

H18 (CTD) 8627 24609 4216 10064 30328 0 59964

H19 (CTD) 7947 17425 2584 8432 24616 0 40664

H20 (Cont.) 9945 18190 5712 10200 25432 0 34000

H21 (TMX) 8585 17723 2176 8024 20808 0 34544

UK22 (TMX) 3550 5050 1280 2080 7920 96 8080

UK23 (Cont.) 2855 2330 1280 1440 4240 0 3600

UK24 (CTD) 3940 1025 160 0 4160 0 1840

UK25 (Cont.) 3675 5525 1600 3280 8640 80 5760

UK26 (TMX) 3600 2200 1680 1760 4080 0 4720

UK27 (CTD) 3500 920 1280 208 3520 0 3008

UK28 (CTD) 2500 3500 1040 2000 5840 0 4320

UK29 (Cont.) 2370 6450 1680 2560 11200 0 8000

UK30 (TMX) 1405 6800 2880 3520 12320 0 9040

15

UK31 (Cont.) 2245 2375 960 640 2240 0 8480

UK32 (TMX) 2025 2375 976 640 1760 0 9200

UK33 (CTD) 1800 2885 1552 2080 4480 160 4960

16

Table S2B

Site (treatment)

Worker (overwinter)

Combined brood

(overwinter)

Storage cells (overwinter)

Number of surviving hives (out of 6)

following winter

G1 (Cont.) 3926 400 12880 6 G2 (CTD) 3796 960 15440 6 G3 (TMX) 2314 880 11200 5 G4 (Cont.) 3588 240 16640 6 G5 (CTD) 4784 800 14640 6 G6 (TMX) 3432 80 14640 4 G7 (Cont.) 4030 720 19520 6 G8 (CTD) 4810 720 12800 5 G9 (TMX) 6474 1200 13440 5 H10 (Cont.) 8840 1360 38760 5 H11 (TMX) 10625 5440 30328 6 H12 (CTD) 5398 952 40256 6 H13 (Cont.) 6545 2312 25568 6 H14 (TMX) 6035 1224 25160 6 H15 (CTD) 5653 1224 24888 6 H16 (Cont.) 10710 3400 33728 6 H17 (TMX) 10073 2584 39440 6 H18 (CTD) 7863 2312 33728 5 H19 (CTD) 5270 2312 27880 6 H20 (Cont.) 7055 2176 22712 6 H21 (TMX) 8628 2856 25568 5 UK22 (TMX) 0 0 0 1 UK23 (Cont.) 375 480 6480 3 UK24 (CTD) 0 0 0 0 UK25 (Cont.) 0 0 0 1 UK26 (TMX) 0 0 0 2 UK27 (CTD) 0 0 0 1 UK28 (CTD) 0 0 0 2 UK29 (Cont.) 850 320 12560 4 UK30 (TMX) 1925 1440 7520 4 UK31 (Cont.) 0 0 0 2 UK32 (TMX) 0 0 0 1 UK33 (CTD) 0 0 0 2

17

Table S2C

Site (treatment)

Bombus workers pre-exposure

(based on multihives of 3

colonies)

Bombus Queen

production

Bombus drone

production

Bombus worker number

Bombus peak

multihive weight (kg)

Osmia reproductive cell number

G1 (Cont.) 275 387 138 133 1.962 80

G2 (CTD) 321 299 253 141 2.461 164

G3 (TMX) 270 368 228 281 2.297 82

G4 (Cont.) 315 157 122 217 1.551 93

G5 (CTD) 349 318 129 129 1.355 59

G6 (TMX) 309 189 149 256 1.268 140

G7 (Cont.) 323 363 178 240 1.95 197

G8 (CTD) 298 368 123 197 2.611 38

G9 (TMX) 332 296 519 292 2.177 284

H10 (Cont.) 240 441 168 223 2.161 116

H11 (TMX) 248 483 163 225 1.693 82

H12 (CTD) 230 294 178 107 1.788 42

H13 (Cont.) 278 404 166 370 2.085 189

H14 (TMX) 268 513 150 230 1.986 286

H15 (CTD) 248 480 214 301 1.301 381

H16 (Cont.) 213 606 167 204 1.779 171

H17 (TMX) 230 312 104 122 1.731 177

H18 (CTD) 300 463 188 295 2.088 97

H19 (CTD) 230 322 94 131 2.091 11

H20 (Cont.) 243 345 164 105 1.72 60

H21 (TMX) 238 382 230 142 2.024 183

UK22 (TMX) 267 219 146 96 2.639 0 †

UK23 (Cont.) 238 165 505 394 2.934 0 †

UK24 (CTD) 276 278 257 210 2.192 20

UK25 (Cont.) 232 82 245 154 1.636 36

UK26 (TMX) 280 277 315 299 1.865 264

UK27 (CTD) 287 315 344 384 2.083 0 †

UK28 (CTD) 295 174 369 301 1.774 45

UK29 (Cont.) 300 305 522 422 2.612 86

UK30 (TMX) 282 63 295 305 1.916 147

UK31 (Cont.) 304 139 109 116 1.251 26

UK32 (TMX) 270 265 105 144 1.549 100

UK33 (CTD) 287 249 144 258 1.326 83 †As no O. bicornis cells were found at these sites no residue analysis for neonicotinoids was possible.

18

Table S3. Statistical tests and analysis of pollen and nectar samples. Summary of statistical tests (A) and results from statistical analyses testing the probability that pollen and nectar samples collected from the oilseed rape crop will express clothianidin, thiamethoxam and imidacloprid above the LoQ (0.53 ng g–1 w/w) in response to seed treatment (B). Due to the relatively infrequent recovery of residues from the crop, separate analyses for pollen and nectar were not undertaken as nonzero data was too scarce for a robust analysis. In these analyses, we used likelihood ratio tests to assess whether (i) there was an overall seed treatment effect common to all countries (TEST 1), (ii) if there was an additive seed treatment effect that was unique to each country (TEST 2), and (iii) if there was an additive seed treatment effect was that a better fit to the data that a simple overall seed treatment effect (TEST 3). TEST 1 compares a seed treatments only model (y ~ treatment + block/country) to the null model (‘y ~ block/country’); TEST 2 compares an additive seed treatment model (y ~ treatment*country + block/country) to the same null model (y ~ block/country); and TEST 3 compares the additive seed treatment model (y ~ treatment*country + block/country) to the simple seed treatment only model (y ~ treatment + block/country). This final comparison is only pertinent when the additive seed treatment model explains more variation than the null model. For both thiamethoxam and imidacloprid only three sites were identified where residues were found in the crop (Table S12).

19

Table S3A Test Description of models compared using likelihood ratio

tests

TEST 1: ‘Seed treatment only model’ > null

H0: y ~ block/country H1: y ~ treatment + block/country

TEST 2: ‘Seed treatment × country model’ > null

H0: y ~ block/country H1: y ~ treatment*country + block/country

TEST 3: ‘Seed treatment × country model’ > ‘Seed treatment only model’

H0: y ~ treatment + block/country H1: y ~ treatment*country + block/country

Table S3B

Analysis Proportion of pollen and nectar samples with

neonicotinoid residues above the LoQ (>0.53 ng g–1 w/w) Clothianidin Thiamethoxam Imidacloprid

Error distribution Binomial (logit Link)

Binomial (logit Link)

Binomial (logit Link)


χ22=6.46,

p=0.04 χ2

2=2.98, p=0.22 χ22=0.19, p=0.90


χ26=12.4,

p=0.05 χ2

6=2.98, p=0.81 χ26=0.19, p=0.99


χ24=5.92,

p=0.20 χ2

4<0.001, p=0.999

χ24<0.001, p=0.999

20

Table S4. Statistical tests and analysis of honey bees for seed treatment effects. Summary of statistical tests (A) and results from statistical analyses for honey bees during the oilseed rape flowering (B) and overwintering periods (C). In these analyses we used likelihood ratio test to assess whether (i) there was an overall seed treatment effect common to all countries (TEST 1), (ii) if there was an additive seed treatment effect that was unique to each country (TEST 2), and (iii) if there was an additive seed treatment effect was that a better fit to the data that a simple overall seed treatment effect (TEST 3). TEST 1 compares a seed treatments only model (y ~ treatment + covariates + block/country) to the null model (‘y ~ covariates + block/country’); TEST 2 compares an additive seed treatment model (y ~ treatment*country + covariates +block/country) to the same null model (y ~ covariates + block/country); and TEST 3 compares the additive seed treatment model (y ~ treatment*country + covariates + block/country) to the simple seed treatment only model (y ~ treatment + covariates + block/country). The inclusion of covariates in models for TESTS 1 to 3 occurred only where they were demonstrated to explain additional variance in the population metrics over that explained by underlying between country variation. This was assessed using likelihood ratio tests to compare the effect of individual covariate (y~ covariates + country) to a simple country only model (y ~ country). Continuous covariates were percentage cover of oilseed rape, percentage arable cover and the natural logs of maximum (NNIMax) concentrations of neonicotinoid residues (combined clothianidin, thiamethoxam and imidacloprid) in hives and detected from the oilseed rape crop. Note the median values of neonicotinoid expression were not used as they were too data poor to provide viable covariates (Table S12). Also as the UK honey bee colonies were so different in size from those of Germany and Hungary their inclusion in one model violated distributional assumptions and so UK sites were tested separately using a simple additive model (y ~ treatment + block) compared to a null model (y ~ block). Honey bee male brood data was too data poor for a valid statistical analysis and so this is not presented. Individual life stages of overwintering brood were infrequent following the overwintering period and so were summed to provide total overwintering brood (eggs + larvae + pupae + male brood). For count data the underlying distributions were sometimes more optimally modelled using a normal distribution with identify link. We provide a measure of explained variance where a fixed effect was found to improve model fit relative to the null model based on how much of the variation in the residuals had been removed (Expl. Var.). Where significant seed treatment differences were identified predicted marginal means were used to compare within treatment differences between the control and thiamethoxam (C-TMX) and control and clothianidin (C-CTD) seed treatments. Note that for the UK overwintering mortality was very high across all seed treatments. Too few data were therefore present in the UK for a valid statistical analysis.

21


tests.


H0: y ~ covariates† + block/country H1: y ~ treatment + covariates† + block/country


H0: y ~ covariates† + block/country H1: y ~ treatment*country + covariates† + block/country


H0: y ~ treatment + covariates† + block/country H1: y ~ treatment*country + covariates† + block/country

Oilseed rape % cover H0: y ~ country H1: y ~ arable_cover + country

Arable crop % cover H0: y ~ country H1: y ~ oilseed_cover + country

Loge (NNIMax hives) H0: y ~ country H1: y ~ NNIMax hives + country

Loge (NNIMax Crop) H0: y ~ country H1: y ~ NNIMax Crop + country

† Covariates are only included in these models where they are identifies as explaining a significant increase in mode variance above that of a simple country only model.

22

Table S4B Honey bee oilseed rape flowering period analysis

Worker (max.

flowering period)

Eggs (max. flowering period)

Larvae (max.

flowering period)†

Pupae (max.

flowering period)

Stored hive products

(max. summed

nectar and pollen cells)

Germany & Hungary analysis Error distribution Neg. Bin.

(link-Log) Normal (Link=Ident.)

Neg. Bin. (link-Log)



Shapiro-Wilk normality test

W=0.87, p=0.001

W=0.94, p=0.11

W=0.90, p<0.01

W=0.88, p<0.01

W=0.89, p<0.01


χ22=6.24,

p=0.04 χ2

2=0.16, p=0.81

χ22=0.06,

p=0.96 χ2

2=0.62, p=0.73

χ22=6.46,

p=0.04


χ26=16.6,

p=0.01 (Expl.Var. = 45.3 %)

Ger: C-TMX z=-0.27, p=0.95; C-CTD z=-0.03, p=0.99. Hun: C-TMX z=-0.45, p=0.89; C-CTD z=0.39, p=0.91. UK: C-TMX z=0.19, p=0.97; C-CTD z=4.38, p<0.01;

χ26=4.13,

p=0.01 (Expl.Var. = 49.9 %)

Ger: C-TMX z=-2.77, p=0.01; C-CTD z=-1.96, p=0.12. Hun: C-TMX z=2.43, p=0.04; C-CTD z=2.12, p=0.08. UK: C-TMX z=-0.42, p=0.90; C-CTD z=0.48, p=0.87.

χ26=2.31,

p=0.88 χ2

6=5.73, p=0.45

χ26=40.5,

p<0.001 (Expl.Var. = 53.6%)

Ger: C-TMX z=0.79, p=0.70; C-CTD z=0.54, p=0.85. Hun: C-TMX z=-1.36, p=0.36; C-CTD z=-1.76, p=0.18. UK: C-TMX z=-2.35, p=0.05; C-CTD z=6.41, p<0.01.


χ24=10.3,

p=0.04 χ2

4=3.97, p=0.004

χ24=2.25,

p=0.68 χ2

4=5.11, p=0.27

χ24=34.1,

p<0.001

Oilseed rape % cover χ21=1.42,

p=0.23 χ2

1=0.01, p=0.84

χ21=1.11,

p=0.29 χ2

1=1.78, p=0.18

χ21=0.02,

p=0.87 Arable crop % cover χ2

1=0.01, p=0.93

χ21=0.21,

p=0.48 χ2

1=1.08, p=0.29

χ21=1.64,

p=0.19 χ2

1=2.987, p=0.09

Loge (NNIMax hives) χ21=0.04,

p=0.83 χ2

1=0.47, p=0.28

χ21=0.23,

p=0.62 χ2

1=0.06, p=0.79

χ21=0.31,

p=0.57 Loge (NNIMax Crop) χ2

1=0.19, p=0.65

χ21=0.03,

p=0.78 χ2

1=0.56, p=0.45

χ21=0.12,

p=0.91 χ2

1=0.37, p=0.54

†Model convergence for larval data could not be achieved where an outlier zero count value was included in a clothianidin seed treatment in the UK. This data point was excluded.

23

Table S4C Honey bee overwintering period analysis

Overwintering hive survival

Overwintering Worker numbers

Overwintering total brood

(eggs + larvae + pupae + male

brood)

Overwintering stored hive products.

Germany & Hungary analysis Error distribution Binomial

(Link=Logit.) Normal (link=Ident.)


Normal (link=Ident.)


NA W=0.94, p=0.24 W=0.87, p=0.01 W=0.91, p=0.08


χ22=3.39,

p=0.18 χ2

2=0.61, p=0.13 χ22=0.93,

p=0.62 χ2

2=0.05, p=0.63


χ26=9.31,

p=0.15 χ2

6=1.47, p<0.01 (Expl.Var. = 59.4 %)

Ger.:C-CTD z=-0.67, p=0.77; C-TMX=-0.25, p=0.99. Hun: C-CTD z=2.85, p=0.01; C-TMX-0.70, p=0.76. UK: NA

χ26=7.59,

p=0.10 χ2

6=0.12, p=0.72


χ24=5.91,

p=0.21 χ2

4=0.85, p=0.01 χ24=6.65,

p=0.03 χ2

4=0.07, p=0.55


p=0.58 χ2

1=0.001, p=0.99 χ21=1.22,

p=0.26 χ2

1=0.08, p=0.44

Arable % cover χ21=2.11,

p=0.14 χ2

1=0.01, p=0.89 χ21=0.27,

p=0.60 χ2

1=0.20, p=0.23

Loge (NNIMax Hives) χ21=1.00,

p=0.31 χ2

1=0.10, p=0.51 χ21=0.02,

p=0.88 χ2

1=0.37, p=0.10

Loge (NNIMax Crop) χ21=3.27,

p=0.08 χ2

1=0.001, p=0.96 χ21=0.34,

p=0.55 χ2

1=0.24, p=0.18

24

Table S5. Statistical tests and analysis of wild bee seed treatment effects. Summary of statistical tests (A) and results from statistical analyses for wild bees (B. terrestris and O. bicornis) during the oilseed rape flowering (B). In these analyses we used likelihood ratio test to assess whether (i) there was an overall seed treatment effect common to all countries (TEST 1), (ii) if there was an additive seed treatment effect that was unique to each country (TEST 2), and (iii) if there was an additive seed treatment effect was that a better fit to the data that a simple overall seed treatment effect (TEST 3). TEST 1 compares a seed treatments only model (y ~ treatment + covariates + block/country) to the null model (‘y ~ covariates + block/country’); TEST 2 compares an additive seed treatment model (y ~ treatment*country + covariates + block/country) to the same null model (y ~ covariates + block/country); and TEST 3 compares the additive seed treatment model (y ~ treatment*country + covariates + block/country) to the simple seed treatment only model (y ~ treatment + covariates + block/country). The inclusion of covariates in models for TESTS 1 to 3 occurred only where they were demonstrated to explain additional variance in the population metrics over that explained by underlying between country variations. This was assessed using likelihood ratio tests to compare the effect of individual covariate (y~ covariates + country) to a simple country only model (y~ country). Continuous covariates were percentage cover of oilseed rape, percentage arable cover and the natural logs of median (NNImedian) and maximum (NNIMax) concentrations of neonicotinoid residues (combined clothianidin, thiamethoxam and imidacloprid) in nests and detected from the oilseed rape crop. Note the median values of neonicotinoid expression in the crop were not used as there were too few data to provide viable covariates (Table S12). For O. bicornis reproductive cell production neonicotinoid residues could only be collected from sites where cells were found and so there was no data for sites UK22, UK23 and UK27. Although significant effects of NNImedian were identified for O. bicornis reproductive cell production the inclusion of these covariates in the additive treatment and country specific treatment models caused the loss of zero count data from three sites. Both seed treatment models were therefore tested without the inclusion of NNImedian so that zero count data was not lost. We provide a measure of explained variance where a fixed effect was found to improve model fit relative to the null model based on how much of the variation in the residuals had been removed (Expl. Var.). Where significant seed treatment differences were identified predicted marginal means to compare within treatment differences between the control and thiamethoxam (C-TMX) and control and clothianidin (C-CTD) seed treatments.

25


tests.


H0: y ~ covariates† + block/country H1: y ~ treatment + covariates† + block/country


H0: y ~ covariates† + block/country H1: y ~ treatment*country + covariates† + block/country


H0: y ~ treatment + covariates† + block/country H1: y ~ treatment*country + covariates† + block/country

Oilseed rape % cover H0: y ~ country H1: y ~ arable_cover + country

Arable crop % cover H0: y ~ country H1: y ~ oilseed_cover + country

Loge (NNIMedian hives) H0: y ~ country H1: y ~ NNIMedian hives + country

Loge (NNIMax hives) H0: y ~ country H1: y ~ NNIMax hives + country

Loge (NNIMax Crop) H0: y ~ country H1: y ~ NNIMax Crop + country

†Covariates are only included in these models where they are identified as explaining a significant increase in mode variance above that of a simple country only model.

26

Table S5B Wild pollinator responses

B. terrestris worker number

B. terrestris

queen number

B. terrestris drone

number

Max. B. terrestris colony

weight gain

O. bicornis reproductive cell number

Error distribution Normal (Ident. Link)

Normal (Ident. Link)


Normal (Ident. link)



W=0.98, p=0.94

W=0.93, p=0.96

W=0.81, p<0.001

W=0.96, p=0.45

W=0.90, p=0.01


χ22=0.19,

p=0.89 χ2

2=0.04, p=0.96

χ22=0.06,

p=0.96 χ2

2=0.09, p=0.91

χ22=4.12,

p=0.12


χ26=4.69,

p=0.46 χ2

6=0.97, p=0.95

χ26=13.1,

p=0.04 (Expl.Var. = 13.6 %)

Ger: C-TMX z=-0.-3.11, p=0.01; C-CTD z=-0.74, p=0.73. Hun: C-TMX z=0.18, p=0.98; C-CTD z=-0.05, p=0.99. UK: C-TMX z=2.41, p=0.04; C-CTD z=0.87, p=0.65.

χ26=1.86,

p=0.74

χ26=7.42,

p=0.28


χ24=4.50,

p=0.25 χ2

4=0.94, p=0.81

χ24=13.1,

p=0.01 χ2

4=1.76, p=0.50

χ24=3.30,

p=0.50


p=0.61 χ2

1=0.14, p=0.59

χ21=0.24,

p=0.62 χ2

1=1.34, p=0.25

χ21=0.60,

p=0.43 Arable % cover χ2

1=0.39 p=0.53

χ21=0.14,

p=0.60 χ2

1=0.14, p=0.70

χ21=1.32,

p=0.26 χ2

1=1.34, p=0.24

Loge (NNImedian Nests) χ21=1.53,

p=0.21 χ2

1=0.15, p=0.59

χ21=1.92,

p=0.16 χ2

1=0.80, p=0.38

χ21=4.33,

p=0.04‡

(Expl.Var. = 0.8 %)

Loge (NNIMax Nests) χ21=0.08,

p=0.77 χ1=2.09, p=0.03†

(Expl.Var. = 13.5 %)

χ21=0.15,

p=0.69 χ2

1=0.71, p=0.41

χ21=0.45,

p=0.49

Loge (NNIMax Crop) χ21=1.50,

p=0.21 χ2

1=0.01, p=0.91

χ21=0.90,

p=0.34 χ2

1=0.81, p=0.37

χ21=0.06,

p=0.79 †If sites where the imidacloprid contaminant are excluded this correlation remains significant (χ2

1=2.14, p=0.02). ‡If sites where the imidacloprid contaminant are excluded this correlation is not significant (χ2

1=0.05, p=0.81). ††This marginal (p=0.07) response of drone production to combined neonicotinoid concentrations in expressed I the oilseed rape crop had a correlation coefficient of –3.14.

27

Table S6. Statistical test and analysis of the exposure of bees to neonicotinoids. Summary of statistical tests (A) and results from statistical analyses for honey bees and wild bees (B. terrestris and O. bicornis) describing the NNIMax index (combining clothianidin, thiamethoxam and imidacloprid peak values) found in stored hive products (pollen and nectar) during the oilseed rape flowering (B). The NNIMax index provides a measure of the exposure of bees to neonicotinoids originating from multiple sources (crop and non-crop plants). Note that assessment of separate neonicotinoid products was not viable due to paucity of data (Table S7 to S9). In these analyses we used likelihood ratio test to assess whether (i) there was an overall seed treatment effect common to all countries (TEST 1), (ii) if there was an additive seed treatment effect that was unique to each country (TEST 2), and (iii) if there was an additive seed treatment effect was that a better fit to the data that a simple overall seed treatment effect (TEST 3). TEST 1 compares a seed treatments only model (y ~ treatment + block/country) to the null model (y ~ block/country); TEST 2 compares an additive seed treatment model (y ~ treatment*country + block/country) to the same null model (y ~ block/country); and TEST 3 compares the additive seed treatment model (y ~ treatment*country + block/country) to the simple seed treatment only model (y ~ treatment + block/country). This final comparison is only pertinent when the additive seed treatment model explains more variation than the null model. For O. bicornis reproductive cell production neonicotinoid residues could only be collected from sites where cells were found and so there was no data for sites UK22, UK23 and UK27.

28

Table S6A

Test Description of models compared using likelihood ratio tests.


H0: y ~ block/country H1: y ~ treatment + block/country


H0: y ~ block/country H1: y ~ treatment*country + block/country


H0: y ~ treatment + block/country H1: y ~ treatment*country + block/country

Table S6B

Expression of neonicotinoids in stored hive products

Honey bee NNIMAX for stored hive products

B. terrestris NNIMAX for stored hive products

O. bicornis NNIMAX for stored hive products

Error distribution Normal (Ident. Link)




χ22=2.78,

p=0.20 χ2

2=0.83, p=0.58

χ22=1.10,

p=0.55


χ26=6.99,

p=0.20 χ2

6=2.74, p=0.79

χ26=6.57,

p=0.22


χ24=4.21,

p=0.27 χ2

4=1.90, p=0.70

χ24=5.46,

p=0.14

29

Table S7.

Concentration of clothianidin, thiamethoxam and imidacloprid residues found in combined pollen and nectar samples collected by honey bees (Table S7a) as well as the same values for residues originating from pollen only (Table S7b) and nectar only samples (Table S7c). Values below the limit of quantification (0.53 ng g–1 wet weight) were set to half the limit of detection (LoD=0.38 ng g–1, Non-detect=0.19 ng g–1). . The combined index of all neonicotinoid residues used as covariates in analyses is given for overall median (NNImedian) and maximum (NNIMax) which account for differences in toxicity of the three neonicotinoid products by weighting them on the basis of acute honey bee oral LD50 values (TMX=0.005 μg bee–1; CTD= 0.00379 μg bee–1; IMI=0.0037 μg bee–1) (29, 30, 31). The NNImedian and NNIMax values were derived for two separate periods: 1) the oilseed rape flowering period used as covariates (pre.); 2) the entire year including pollen and nectar samples collected in the winter period (post). The latter of these was used as a covariate for the overwintering colony strength assessments. Note NNIMax values were the same for both periods. Table S7A: Combined pollen and nectar Site (treat.)

Clothianidin (ng g–1 w/w)

Thiamethoxam (ng g–1 w/w)

Imidacloprid (ng g–1 w/w)

NNImedinan (ng g–1 w/w)

NNIMax (ng g–1 w/w) Min. Med Max Min Med Max Min Med Max Pre. Post

G1 (Cont.) 0.19 0.19 1.2 0.19 0.19 0.19 0.19 0.19 0.19 0.48 0.48 1.24

G2 (CTD) 0.19 0.19 0.59 0.19 0.19 0.19 0.19 0.19 0.87 0.48 0.48 1.28

G3 (TMX) 0.19 0.19 0.74 0.19 0.19 0.19 0.19 0.19 0.19 0.48 0.48 0.89

G4 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.58 0.19 0.19 0.19 0.48 0.48 0.86

G5 (CTD) 0.19 0.67 1.19 0.19 0.19 0.19 0.19 0.19 0.19 0.84 0.83 1.24

G6 (TMX) 0.19 0.19 0.70 0.19 0.19 0.65 0.19 0.19 0.19 0.48 0.48 1.32

G7 (Cont.) 0.19 0.19 0.77 0.19 0.19 0.19 0.19 0.19 0.19 0.48 0.48 0.92

G8 (CTD) 0.19 0.73 1.21 0.19 0.19 0.19 0.19 0.19 0.19 0.89 0.8 1.25

G9 (TMX) 0.19 0.19 0.68 0.19 0.46 1.41 0.19 0.19 1.00 0.74 0.48 2.65

H10 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 1.77 0.48 0.48 1.65

H11 (TMX) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.48 0.48 0.48

H12 (CTD) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.48 0.48 0.48

H13 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.73 0.19 0.19 0.19 0.48 0.48 1.01

H14 (TMX) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.48 0.48 0.48

H15 (CTD) 0.19 0.19 1.05 0.19 0.19 0.19 0.19 0.19 0.19 0.48 0.48 1.13

H16 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.48 0.48 0.48

H17 (TMX) 0.19 0.19 0.19 0.19 0.19 0.79 0.19 0.19 0.19 0.48 0.48 1.07

H18 (CTD) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.48 0.48 0.48

H19 (CTD) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.48 0.48 0.48

H20 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 5.99 0.48 0.48 4.77

H21 (TMX) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 3.10 0.48 0.48 2.63

UK22 (TMX) 0.19 0.19 0.19 0.19 0.19 0.86 0.19 0.19 0.19 0.48 0.48 1.14

UK23 (Cont.) 0.19 0.19 0.77 0.19 0.19 0.40 0.19 0.19 0.19 0.48 0.48 1.12

30

UK24 (CTD) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.48 0.48 0.48

UK25 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.48 0.48 0.48

UK26 (TMX) 0.19 0.62 1.35 0.19 0.19 0.63 0.19 0.19 0.19 0.8 0.48 1.79

UK27 (CTD) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.48 0.48 0.48

UK28 (CTD) 0.19 0.19 0.78 0.19 0.19 0.19 0.19 0.19 0.19 0.48 0.48 0.92

UK29 (Cont.) 0.19 0.56 0.96 0.19 0.19 0.19 0.19 0.19 0.19 0.75 0.48 1.06

UK30 (TMX) 0.19 0.19 1.48 0.19 0.19 0.58 0.19 0.19 0.19 0.48 0.48 1.83

UK31 (Cont.) 0.19 0.7 1.46 0.19 0.19 0.19 0.19 0.19 0.19 0.86 0.48 1.44

UK32 (TMX) 0.19 0.19 0.66 0.19 0.19 0.19 0.19 0.19 0.19 0.48 0.48 0.83

UK33 (CTD) 0.19 0.19 0.99 0.19 0.19 0.19 0.19 0.19 0.19 0.48 0.48 1.08

31

Table S7B: Pollen only

Site (Treat.) Clothianidin (ng g–1 w/w)



Min. Med Max Min. Med Max Min Med Max

G1 (Cont.) 0.19 0.38 0.59 0.19 0.19 0.19 0.19 0.19 0.19

G2 (CTD) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.87

G3 (TMX) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

G4 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.58 0.19 0.19 0.19

G5 (CTD) 0.6 0.78 1.08 0.19 0.19 0.19 0.19 0.19 0.19

G6 (TMX) 0.19 0.19 0.19 0.19 0.19 0.57 0.19 0.19 0.19

G7 (Cont.) 0.19 0.19 0.77 0.19 0.19 0.19 0.19 0.19 0.19

G8 (CTD) 0.19 0.5 0.73 0.19 0.19 0.19 0.19 0.19 0.19

G9 (TMX) 0.19 0.19 0.19 0.19 0.33 0.67 0.19 0.19 0.19

H10 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

H11 (TMX) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

H12 (CTD) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

H13 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

H14 (TMX) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

H15 (CTD) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

H16 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

H17 (TMX) 0.19 0.19 0.19 0.19 0.19 0.71 0.19 0.19 0.19

H18 (CTD) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

H19 (CTD) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

H20 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 5.99

H21 (TMX) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

UK22 (TMX) 0.19 0.19 0.19 0.19 0.19 0.86 0.19 0.19 0.19

UK23 (Cont.) 0.19 0.19 0.77 0.19 0.19 0.40 0.19 0.19 0.19

UK24 (CTD) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

UK25 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

UK26 (TMX) 0.19 0.41 0.88 0.19 0.19 0.19 0.19 0.19 0.19

UK27 (CTD) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

UK28 (CTD) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

UK29 (Cont.) 0.19 0.19 0.56 0.19 0.19 0.19 0.19 0.19 0.19

UK30 (TMX) 0.19 0.19 0.19 0.19 0.37 0.58 0.19 0.19 0.19

UK31 (Cont.) 0.19 0.45 0.92 0.19 0.19 0.19 0.19 0.19 0.19

UK32 (TMX) 0.19 0.19 0.66 0.19 0.19 0.19 0.19 0.19 0.19

UK33 (CTD) 0.19 0.19 0.99 0.19 0.19 0.19 0.19 0.19 0.19

32

Table S7B: Nectar only

Site (Treat.) Clothianidin (ng g–1 w/w)




G1 (Cont.) 0.19 0.19 1.2 0.19 0.19 0.19 0.19 0.19 0.19

G2 (CTD) 0.19 0.57 0.59 0.19 0.19 0.19 0.19 0.19 0.19

G3 (TMX) 0.19 0.19 0.74 0.19 0.19 0.19 0.19 0.19 0.19

G4 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

G5 (CTD) 0.19 0.67 1.19 0.19 0.19 0.19 0.19 0.19 0.19

G6 (TMX) 0.19 0.19 0.7 0.19 0.19 0.65 0.19 0.19 0.19

G7 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

G8 (CTD) 0.84 0.98 1.21 0.19 0.19 0.19 0.19 0.19 0.19

G9 (TMX) 0.19 0.19 0.68 0.19 0.65 1.41 0.19 0.74 1

H10 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 1.77

H11 (TMX) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

H12 (CTD) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

H13 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.73 0.19 0.19 0.19

H14 (TMX) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

H15 (CTD) 0.19 0.19 1.05 0.19 0.19 0.19 0.19 0.19 0.19

H16 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

H17 (TMX) 0.19 0.19 0.19 0.19 0.67 0.79 0.19 0.19 0.19

H18 (CTD) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

H19 (CTD) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

H20 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

H21 (TMX) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 3.1

UK22 (TMX) 0.19 0.19 0.19 0.19 0.19 0.64 0.19 0.19 0.19

UK23 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

UK24 (CTD) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

UK25 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

UK26 (TMX) 0.19 0.88 1.35 0.19 0.63 0.63 0.19 0.19 0.19

UK27 (CTD) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

UK28 (CTD) 0.19 0.49 0.78 0.19 0.19 0.19 0.19 0.19 0.19

UK29 (Cont.) 0.69 0.83 0.96 0.19 0.19 0.19 0.19 0.19 0.19

UK30 (TMX) 0.75 1.19 1.48 0.19 0.19 0.19 0.19 0.19 0.19

UK31 (Cont.) 0.19 1.45 1.46 0.19 0.19 0.19 0.19 0.19 0.19

UK32 (TMX) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

UK33 (CTD) 0.19 0.63 0.95 0.19 0.19 0.19 0.19 0.19 0.19

33

Table S8. Concentration of clothianidin, thiamethoxam and imidacloprid residues in pollen and nectar. Concentration of clothianidin, thiamethoxam and imidacloprid residues found in combined pollen and nectar samples collected by B. terrestris (A) as well as the same values for pollen only (B) and nectar only (C). Values below the limit of quantification (0.53 ng g–1 wet weight) were set to half the limit of detection (LoD=0.38 ng g–1, Non-detect=0.19 ng g–1). The combined index of all neonicotinoid residues used as covariates in analyses is given for overall median (NNImedian) and maximum (NNIMax) which account for differences in toxicity of the three neonicotinoid products by weighting them on the basis of acute Apis mellifera oral LD50 values. Table S8A: Pollen and nectar combined

Clothianidin (ng g–1 w/w)


Imidacloprid NNImedi

an (ng g–1 w/w)

NNIMax (ng g–1 w/w)


G1 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.48 0.48

G2 (CTD) 0.19 0.19 0.58 0.19 0.19 0.19 0.19 0.19 0.19 0.48 0.77

G3 (TMX) 0.19 0.19 0.19 0.19 0.19 0.93 0.19 0.19 0.19 0.48 1.22

G4 (Cont.) 0.19 0.19 0.19 0.19 0.19 5.05 0.19 0.19 0.19 0.48 5.33

G5 (CTD) 0.19 0.19 0.19 0.19 0.19 1.16 0.19 0.19 0.19 0.48 1.44

G6 (TMX) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.54 0.48 0.74

G7 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.48 0.48

G8 (CTD) 0.19 1.77 2.27 0.19 0.19 0.19 0.19 0.19 0.19 1.68 2.06

G9 (TMX) 0.19 0.19 0.99 0.19 0.76 1.25 0.19 0.19 0.6 1.04 2.43

H10 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.48 0.48

H11 (TMX) 0.19 0.19 1.38 0.19 0.19 0.19 0.19 0.19 0.19 0.48 1.38

H12 (CTD) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.48 0.48

H13 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.48 0.48

H14 (TMX) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.48 0.48

H15 (CTD) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.48 0.48

H16 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.48 0.48

H17 (TMX) 0.19 0.19 0.19 0.19 0.19 1.01 0.19 0.19 0.19 0.48 1.29

H18 (CTD) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.48 0.48

H19 (CTD) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 1.45 0.48 1.41

H20 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.48 0.48

H21 (TMX) 0.19 0.19 0.19 0.19 0.19 0.61 0.19 0.19 0.19 0.48 0.89

UK22 (TMX) 0.19 0.19 0.58 0.19 0.19 0.19 0.19 0.19 0.19 0.48 0.77

UK23 (Cont.) 0.19 1.12 4.69 0.19 0.19 22.2† 0.19 0.19 0.19 1.18 3.89

UK24 (CTD) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.48 0.48

UK25 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 4.07 0.48 3.35

UK26 (TMX) 0.19 1.3 1.85 0.19 0.19 0.76 0.19 0.19 0.19 1.32 2.3

UK27 (CTD) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.48 0.48

UK28 (CTD) 0.19 0.19 1.54 0.19 0.19 0.19 0.19 0.19 1.16 0.48 2.22

34

UK29 (Cont.) 0.19 0.76 5.23 0.19 0.19 1.09 0.19 0.19 0.19 0.9 5.19

UK30 (TMX) 0.19 2.35 2.93 0.19 0.19 0.55 0.19 0.19 0.19 2.12 2.91

UK31 (Cont.) 0.19 1.25 3.05 0.19 0.19 0.82 0.19 0.19 0.19 1.28 3.27

UK32 (TMX) 0.19 0.19 1.71 0.19 0.19 0.64 0.19 0.19 0.19 0.48 2.08

UK33 (CTD) 0.19 0.85 2.07 0.19 0.19 0.19 0.19 0.19 8.70 0.98 8.19

†The identified thiamethoxam residue outlier was ignored in the derivation of the NNIMax.

35

Table S8B Pollen only Site (Treat.) Clothianidin

(ng g–1 w/w) Thiamethoxam

(ng g–1 w/w) Imidacloprid (ng g–1 w/w)


G1 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

G2 (CTD) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

G3 (TMX) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

G4 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

G5 (CTD) 0.19 0.19 0.19 0.19 0.19 1.16 0.19 0.19 0.19

G6 (TMX) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.54

G7 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

G8 (CTD) 0.19 0.19 2.00 0.19 0.19 0.19 0.19 0.19 0.19

G9 (TMX) 0.19 0.19 0.79 0.19 0.19 0.8 0.19 0.19 0.6

H10 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

H11 (TMX) 0.19 0.19 1.38 0.19 0.19 0.19 0.19 0.19 0.19

H12 (CTD) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

H13 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

H14 (TMX) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

H15 (CTD) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

H16 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

H17 (TMX) 0.19 0.19 0.19 0.19 0.19 1.01 0.19 0.19 0.19

H18 (CTD) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

H19 (CTD) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

H20 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

H21 (TMX) 0.19 0.19 0.19 0.19 0.19 0.61 0.19 0.19 0.19

UK22 (TMX) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

UK23 (Cont.) 1.11 1.41 4.69 0.19 0.19 22.2† 0.19 0.19 0.19

UK24 (CTD) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

UK25 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 4.07

UK26 (TMX) 0.19 0.72 1.85 0.19 0.19 0.19 0.19 0.19 0.19

UK27 (CTD) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

UK28 (CTD) 0.19 0.19 1.54 0.19 0.19 0.19 0.19 0.19 1.16

UK29 (Cont.) 0.19 2.63 5.23 0.19 0.19 1.09 0.19 0.19 0.19

UK30 (TMX) 0.19 2.38 2.93 0.19 0.19 0.55 0.19 0.19 0.19

UK31 (Cont.) 0.19 2.87 3.05 0.19 0.19 0.82 0.19 0.19 0.19

UK32 (TMX) 0.19 0.19 1.71 0.19 0.19 0.64 0.19 0.19 0.19

UK33 (CTD) 0.19 0.19 2.07 0.19 0.19 0.19 0.19 0.19 8.69

†The identified thiamethoxam residue outlier was ignored in the derivation of the NNIMax..

36

Table S8C: Nectar only Site (Treat.) Clothianidin




G1 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

G2 (CTD) 0.58 0.58 0.58 0.19 0.19 0.19 0.19 0.19 0.19

G3 (TMX) 0.19 0.19 0.19 0.19 0.56 0.93 0.19 0.19 0.19

G4 (Cont.) 0.19 0.19 0.19 0.19 2.62 5.05 0.19 0.19 0.19

G5 (CTD) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

G6 (TMX) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

G7 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

G8 (CTD) 1.77 2.02 2.27 0.19 0.19 0.19 0.19 0.19 0.19

G9 (TMX) 0.19 0.59 0.99 0.76 1.00 1.25 0.19 0.19 0.19

H10 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

H11 (TMX) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

H12 (CTD) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

H13 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

H14 (TMX) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

H15 (CTD) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

H16 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

H17 (TMX) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

H18 (CTD) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

H19 (CTD) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.82 1.45

H20 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

H21 (TMX) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

UK22 (TMX) 0.53 0.56 0.58 0.19 0.19 0.19 0.19 0.19 0.19

UK23 (Cont.) 0.19 0.19 0.19 0.19 0.43 0.67 0.19 0.19 0.19

UK24 (CTD) 0.12 0.16 0.19 0.19 0.19 0.19 0.19 0.19 0.19

UK25 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

UK26 (TMX) 1.3 1.34 1.37 0.73 0.74 0.76 0.19 0.19 0.19

UK27 (CTD) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

UK28 (CTD) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

UK29 (Cont.) 0.6 0.68 0.76 0.19 0.19 0.19 0.19 0.19 0.19

UK30 (TMX) 1.64 2.00 2.35 0.19 0.37 0.55 0.19 0.19 0.19

UK31 (Cont.) 0.96 1.11 1.25 0.19 0.19 0.19 0.19 0.19 0.19

UK32 (TMX) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

UK33 (CTD) 0.85 0.91 0.96 0.19 0.19 0.19 0.19 0.19 0.19

37

Table S9. Concentration of clothianidin, thiamethoxam and imidacloprid residues found in pollen and nectar. Concentration of clothianidin, thiamethoxam and imidacloprid residues found in combined pollen and nectar samples collected by O. bicornis (A) as well as the same values for pollen only (B) and nectar only (C). Values below the limit of quantification (0.52 ng g–1 wet weight) were set to half the limit of detection (LoD=0.37 ng g–1, Non-detect=0.185 ng g–1). The combined index of all neonicotinoid residues used as covariates in analyses is given for overall median (NNImedian) and maximum (NNIMax) which account for differences in toxicity of the three neonicotinoid products by weighting them on the basis of acute Apis mellifera oral LD50 values. For three sites (UK22, UK23 and UK27) no O. bicornis cells were found. As such it was not possible to determine residue concentrations in pollen and nectar collected by the bee for these sites. Table S9A: Combined pollen and nectar

Site (treat.) Clothianidin (ng g–1 w/w)



Min. Med Max Min Med Max Med Med Max

G1 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

G2 (CTD) 0.19 0.19 0.19 0.19 0.19 1.22 0.19 0.19 0.19

G3 (TMX) 0.19 0.19 1.97 0.19 1.26 3.22 0.19 0.19 0.19

G4 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

G5 (CTD) 0.19 0.19 0.19 0.19 0.19 1.57 0.19 0.19 0.19

G6 (TMX) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

G7 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.65

G8 (CTD) 0.19 0.19 0.19 0.19 1.1 1.42 0.19 0.81 0.81

G9 (TMX) 0.19 0.19 1.07 0.19 0.85 0.96 0.19 0.66 0.70

H10 (Cont.) 0.19 0.49 0.8 0.19 1.17 2.15 0.19 0.19 0.19

H11 (TMX) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 1.78 3.37

H12 (CTD) 0.19 0.19 0.19 0.19 2.47 2.47 0.54 4.01 4.01

H13 (Cont.) 0.19 0.19 0.73 0.19 0.19 0.19 0.19 0.19 2.45

H14 (TMX) 0.19 0.76 0.98 0.19 0.19 0.88 0.19 0.19 0.19

H15 (CTD) 0.19 0.19 0.19 0.19 0.19 0.61 0.19 0.19 0.87

H16 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.6 1.00

H17 (TMX) 0.19 0.19 1.25 0.19 0.19 0.19 0.19 0.19 0.56

H18 (CTD) 0.19 0.19 0.19 0.19 0.19 0.61 0.19 0.19 1.00

H19 (CTD) 0.19 0.19 0.19 0.19 0.19 0.19 8.42 8.43 8.43

H20 (Cont.) 0.19 1.54 1.54 0.19 0.19 0.19 0.19 0.19 0.19

H21 (TMX) 0.19 0.19 1.27 0.19 0.19 0.19 0.19 0.19 0.19

UK22 (TMX) NA† NA† NA† NA† NA† NA† NA† NA† NA†

UK23 (Cont.) NA† NA† NA† NA† NA† NA† NA† NA† NA†

UK24 (CTD) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

UK25 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

38

UK26 (TMX) 0.19 0.19 0.19 0.19 1.11 1.16 0.19 0.19 0.19

UK27 (CTD) NA† NA† NA† NA† NA† NA† NA† NA† NA†

UK28 (CTD) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

UK29 (Cont.) 0.55 0.19 0.69 0.19 1.13 1.77 0.19 0.19 0.19

UK30 (TMX) 0.19 0.19 0.19 0.19 0.45 1.05 0.19 0.19 0.19

UK31 (Cont.) 1.60 0.72 1.24 0.19 1.74 1.87 0.19 0.39 0.59

UK32 (TMX) 0.19 0.48 1.2 0.19 0.19 1.53 0.19 0.19 0.19

UK33 (CTD) 0.98 0.19 0.19 0.19 1.03 1.9 0.19 0.19 0.19

†As no reproductive cells were found at this site it was not possible to collect pollen and nectar samples to assess for residues.

39

Table S9B: Pollen only

Site (treat.) Clothianidin (ng g–1 w/w)



Min. Med Max Min.

Med Max Med Med Max

G1 (Cont.) NA†† NA†† NA†† NA†† NA†† NA†† NA†† NA†† NA††

G2 (CTD) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

G3 (TMX) 1.26 1.26 3.22 0.19 0.19 1.97 0.19 0.19 0.19

G4 (Cont.) NA†† NA†† NA†† NA†† NA†† NA†† NA†† NA†† NA††

G5 (CTD) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

G6 (TMX) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

G7 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

G8 (CTD) 0.19 1.10 1.10 0.19 0.19 0.19 0.19 0.81 0.81

G9 (TMX) 0.96 0.96 0.96 0.19 0.19 0.19 0.7 0.7 0.7

H10 (Cont.) 2.15 2.15 2.15 0.8 0.8 0.8 0.19 0.19 0.19

H11 (TMX) 0.19 0.19 0.19 0.19 0.19 0.19 3.37 3.37 3.37

H12 (CTD) 2.47 2.47 2.47 0.19 0.19 0.19 4.01 4.01 4.01

H13 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

H14 (TMX) 0.19 0.19 0.19 0.98 0.98 0.98 0.19 0.19 0.19

H15 (CTD) 0.61 0.61 0.61 0.19 0.19 0.19 0.19 0.19 0.19

H16 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.19 1.00 1.00 1.00

H17 (TMX) 0.19 0.19 0.19 0.19 1.25 1.25 0.19 0.19 0.19

H18 (CTD) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

H19 (CTD) NA†† NA†† NA†† NA†† NA†† NA†† NA†† NA†† NA††

H20 (Cont.) 0.19 0.19 0.19 1.54 1.54 1.54 0.19 0.19 0.19

H21 (TMX) 0.19 0.19 0.19 0.19 0.73 1.26 0.19 0.19 0.19



UK24 (CTD) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

UK25 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

UK26 (TMX) 1.16 1.16 1.16 0.19 0.19 0.19 0.19 0.19 0.19


UK28 (CTD) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

UK29 (Cont.) 1 1.13 1.26 0.19 0.19 0.19 0.19 0.19 0.19

UK30 (TMX) 0.19 0.62 1.05 0.19 0.19 0.19 0.19 0.19 0.19

UK31 (Cont.) 1.87 1.87 1.87 1.24 1.24 1.24 0.58 0.58 0.58

UK32 (TMX) 0.19 0.86 1.53 0.19 0.69 1.2 0.19 0.19 0.19

UK33 (CTD) 1.03 1.03 1.03 0.19 0.19 0.19 0.19 0.19 0.19

†As no reproductive cells were found at this site it was not possible to collect pollen and nectar samples to assess for residues. †† There was insufficient pollen could be cleanly removed to undertake a viable residue analysis.

40

Table S9C: Nectar only Site (treat.) Clothianidin



Min. Med Max Min Med Max Med Med Max

G1 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

G2 (CTD) 0.85 1.03 1.22 0.19 0.19 0.19 0.19 0.19 0.19

G3 (TMX) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

G4 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

G5 (CTD) 0.19 0.88 1.57 0.19 0.19 0.19 0.19 0.19 0.19

G6 (TMX) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

G7 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.42 0.65

G8 (CTD) 1.42 1.42 1.42 0.19 0.19 0.19 0.81 0.81 0.81

G9 (TMX) 0.19 0.47 0.74 0.19 0.19 1.07 0.19 0.19 0.61

H10 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

H11 (TMX) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

H12 (CTD) 0.19 0.19 0.19 0.19 0.19 0.19 0.55 0.55 0.55

H13 (Cont.) 0.19 0.19 0.19 0.19 0.46 0.73 0.19 1.32 2.45

H14 (TMX) 0.19 0.19 0.88 0.19 0.19 0.53 0.19 0.19 0.19

H15 (CTD) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.87

H16 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

H17 (TMX) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.56

H18 (CTD) 0.19 0.19 0.61 0.19 0.19 0.19 0.19 0.19 1.00

H19 (CTD) 0.19 0.19 0.19 0.19 0.19 0.19 8.43 8.43 8.43

H20 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

H21 (TMX) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19



UK24 (CTD) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

UK25 (Cont.) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

UK26 (TMX) 0.19 0.95 1.07 0.19 0.19 0.19 0.19 0.19 0.19


UK28 (CTD) 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19 0.19

UK29 (Cont.) 0.56 1.3 1.76 0.19 0.19 0.69 0.19 0.19 0.19

UK30 (TMX) 0.19 0.45 0.72 0.19 0.19 0.19 0.19 0.19 0.19

UK31 (Cont.) 1.61 1.61 1.61 0.19 0.19 0.19 0.19 0.19 0.19

UK32 (TMX) 0.19 0.19 0.59 0.19 0.48 1.08 0.19 0.19 0.19

UK33 (CTD) 0.99 1.71 1.9 0.19 0.19 0.19 0.19 0.19 0.19

†As no reproductive cells were found at this site it was not possible to collect pollen and nectar samples to assess for residues.

41

Table S10. Plant species in pollen. Foraging proportion of oilseed rape (OSR pollen) and other plant species (taxon richness) based on microscopy analysis of pollen collected during the oilseed rape flowering period for honey bees, B. terrestris and O. bicornis. Honey bee pollen was collected using pollen traps attached to the front of individual hives on a single sampling date 14 days after the initiation of exposure to the crop. The six samples per site were then amalgamated at the site level. For B. terrestris pollen was removed from returning workers at 14 days after exposure to the oilseed rape. For each B. terrestris multihive (four multihives per site, each containing three colonies) 10 pollen baskets were collected and then amalgamated at the site level. For returning to each hive or colony at a site and then amalgamated. Pollen from O. bicornis was removed from 10 dissected reproductive cells within trap nests. Frequency of oilseed rape in diets was scored along an index of 0-5 (0 = 0%; 0.1=trace amounts; 1 = 0.1-20%; 2=20-40%; 3=40-60%; 4=60-80%; 5= >80%). Pollen was stained using Fuschin and identified under compound light microscope where possible to generic or species level depending on the degree of homogeneity in pollen morphology. For all three species pollen was collected honey bees

Site (treat.) Honey bees B. terrestris O. bicornis OSR

pollen Taxon

richness OSR

pollen Taxon

richness OSR

pollen Taxon

richness

G1 (Cont.) 0.1 6 0.1 4 0.1 3 G2 (CTD) 0.1 5 0.1 6 1.0 4 G3 (TMX) 1.5 4 0.1 6 2.0 4 G4 (Cont.) 1.0 4 0.1 5 0.1 4 G5 (CTD) 2.0 4 0.1 5 0.5 3 G6 (TMX) 0.5 5 0.5 4 2.0 3 G7 (Cont.) 0.1 4 5.0 1 0.1 5 G8 (CTD) 0.5 4 5.0 1 2.0 3 G9 (TMX) 1.0 5 4.5 2 1.0 2 H10 (Cont.) 2.0 5 3.0 2 1.0 3 H11 (TMX) 3.0 4 5.0 1 4.5 3 H12 (CTD) 3.0 3 4.5 2 0.1 2 H13 (Cont.) 1.0 7 5.0 1 2.0 3 H14 (TMX) 1.5 4 2.5 2 4.0 3 H15 (CTD) 5.0 2 5.0 1 0.1 2 H16 (Cont.) 1.5 8 3.0 2 3.5 3 H17 (TMX) 2.0 4 1.5 2 0.5 3 H18 (CTD) 1.0 8 2.0 1 1.0 3 H19 (CTD) 1.0 6 1.0 3 5.0 1 H20 (Cont.) 2.0 7 4.0 2 0.1 3 H21 (TMX) 2.0 6 4.5 1 0.1 3 UK22 (TMX) 1.0 5 2.0 2 NA† NA† UK23 (Cont.) 3.0 4 4.5 2 NA† NA†

42

UK24 (CTD) 3.0 3 1.5 2 2.0 4 UK25 (Cont.) 2.5 4 2.0 2 0.1 3 UK26 (TMX) 3.0 5 5.0 1 0.5 2 UK27 (CTD) 2.0 3 5.0 1 NA† NA† UK28 (CTD) 4.0 2 5.0 1 1.0 2 UK29 (Cont.) 3.0 4 3.5 2 0.1 3 UK30 (TMX) 2.0 3 4.5 2 1.0 3 UK31 (Cont.) 3.0 3 2.0 2 0.0 1 UK32 (TMX) 1.0 2 1.5 2 0.1 4 UK33 (CTD) 2.5 3 0.1 4 1.0 3

†NA, not applicable.

43

Table S11. frequency of Varroa sp. and Nosema sp. infections in honey bees. Mean and ranges for the frequency of Varroa sp. and Nosema sp. infections in honey bee assessed at the end of the oilseed rape flowering period and pre-winter. The proportion of bees with Varroa mites on the body were assessed (based on 100 worker honey bees). For Nosema sp. surface sterilized bees (30 honey bees) were homogenized in 4 mL water and the homogenate was analyzed microscopically (400× magnification). The number of Nosema spp. spores within the visual field were classified on a 4 point semi-quantitative scale where: 0 = no Nosema spores in the field of view; 1 = <10 Nosema spores; 2 = 10-100 Nosema spores; 3 = > 100 spores).

Site (treat.) Oilseed rape flowering period Pre-winter assessment Nosema sp. Varroa sp. (%) Nosema sp. Varroa sp.

G1 (Cont.) 0.5 0.0 0.0 0.7 G2 (CTD) 0.0 0.0 0.0 1.0 G3 (TMX) 2.0 0.0 0.0 1.6 G4 (Cont.) 1.5 0.0 0.0 2.0 G5 (CTD) 0.5 0.4 0.0 1.7 G6 (TMX) 0.0 0.4 0.0 0.8 G7 (Cont.) 0.5 0.0 0.0 0.7 G8 (CTD) 0.5 0.0 0.0 0.7 G9 (TMX) 1.5 0.0 0.0 0.4 H10 (Cont.) 2.0 0.0 0.0 3.0 H11 (TMX) 1.0 0.0 0.5 0.4 H12 (CTD) 1.0 0.7 0.0 6.0 H13 (Cont.) 2.0 0.0 0.0 0.4 H14 (TMX) 1.0 0.0 0.0 0.7 H15 (CTD) 0.5 0.0 0.0 0.4 H16 (Cont.) 0.0 0.0 0.0 0.4 H17 (TMX) 0.0 0.4 0.0 6.4 H18 (CTD) 2.0 0.0 0.0 2.0 H19 (CTD) 2.0 0.0 0.0 2.0 H20 (Cont.) 2.0 0.0 0.0 1.0 H21 (TMX) 2.0 0.0 0.0 3.2 UK22 (TMX) 2.0 0.8 0.0 8.7 UK23 (Cont.) 1.0 2.0 0.0 12.0 UK24 (CTD) 0.0 2.7 NA† NA† UK25 (Cont.) 0.0 1.2 0.0 8.0 UK26 (TMX) 0.0 1.2 0.5 2.0 UK27 (CTD) 0.0 0.4 1.0 4.0 UK28 (CTD) 0.0 0.0 0.0 7.4 UK29 (Cont.) 2.0 0.0 0.0 5.5 UK30 (TMX) 0.0 1.0 0.0 7.2

44

UK31 (Cont.) 0.0 1.6 0.0 12.0 UK32 (TMX) 0.0 1.2 1.0 4.0 UK33 (CTD) 0.0 1.0 0.0 18.0

†NA, not applicable. All hives at site 24 had died by the pre-winter period and so no assessment of disease could be made.

45

Table S12. Concentrations of clothianidin, thiamethoxam and imidacloprid residues in pollen and nectar. Concentrations of clothianidin, thiamethoxam and imidacloprid residues in combined pollen and nectar collected from the oilseed rape crop (A) as well as the same values for pollen only (B) and nectar only (C). Values below the limit of quantification (LoQ= 0.53 ng g–1 wet weight) were set to half the limit of detection (LoD=0.38 ng g–1, Non-detect=0.19 ng g–1). Two pollen and two nectar samples were collected at day 2 and day 9 of crop flowering using honey bees caged over the crop. Pollen was removed from pollen baskets and nectar was dissected from the honey stomachs of worker bees. The median and maximum residues collected from the four pollen and nectar samples from the oilseed rape crop at each site are given, as well as the proportion of pollen and nectar samples above the LoQ (out of a maximum of 4). Table S12A: Pollen and nectar. Site (treat.) Clothianidin



Median (max.)

Prop. Samples > LOQ

Median (max.)

Prop. Samples > LOQ

Median (max.)

Prop. Samples > LOQ

G1 (Cont.) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

G2 (CTD) 0.19 (0.65) 0.25 0.19 (0.19) 0.00 0.19 (0.19) 0.00

G3 (TMX) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

G4 (Cont.) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

G5 (CTD) 0.50 (0.81) 0.50 0.19 (0.19) 0.00 0.19 (0.19) 0.00

G6 (TMX) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

G7 (Cont.) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

G8 (CTD) 0.19 (1.47) 0.25 0.19 (0.19) 0.00 0.19 (0.19) 0.00

G9 (TMX) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

H10 (Cont.) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

H11 (TMX) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (1.20) 0.33

H12 (CTD) 0.19 (2.21) 0.25 0.19 (0.19) 0.00 0.19 (1.10) 0.25

H13 (Cont.) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

H14 (TMX) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

H15 (CTD) 0.19 (0.19) 0.00 0.19 (0.58) 0.25 0.19 (0.19) 0.00

H16 (Cont.) 0.19 (0.19) 0.00 0.19 (0.61) 0.25 0.19 (1.23) 0.25

H17 (TMX) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

H18 (CTD) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

H19 (CTD) 0.19 (0.19) 0.00 0.19 (0.82) 0.25 0.19 (0.19) 0.00

H20 (Cont.) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

H21 (TMX) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

UK22 (TMX) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

UK23 (Cont.) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

46

UK24 (CTD) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

UK25 (Cont.) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

UK26 (TMX) 0.44 (0.76) 0.50 0.19 (0.19) 0.00 0.19 (0.19) 0.00

UK27 (CTD) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

UK28 (CTD) 0.37 (0.55) 0.50 0.19 (0.19) 0.00 0.19 (0.19) 0.00

UK29 (Cont.) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

UK30 (TMX) 0.4 (0.84) 0.25 0.19 (0.19) 0.00 0.19 (0.19) 0.00

UK31 (Cont.) 0.52 (0.87) 0.50 0.19 (0.19) 0.00 0.19 (0.19) 0.00

UK32 (TMX) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

UK33 (CTD) 0.48 (0.79) 0.50 0.19 (0.19) 0.00 0.19 (0.19) 0.00

47

Table S12B: Pollen only Site (treat.) Clothianidin



Median (max.)

Prop. Samples > LOQ

Median (max.)

Prop. Samples > LOQ

Median (max.)

Prop. Samples > LOQ

G1 (Cont.) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

G2 (CTD) 0.42 (0.65) 0.50 0.19 (0.19) 0.00 0.19 (0.19) 0.00

G3 (TMX) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

G4 (Cont.) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

G5 (CTD) 0.81 (0.81) 1.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

G6 (TMX) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

G7 (Cont.) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

G8 (CTD) 0.83 (1.47) 0.50 0.19 (0.19) 0.00 0.19 (0.19) 0.00

G9 (TMX) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

H10 (Cont.) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

H11 (TMX) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.70 (1.2) 0.50

H12 (CTD) 1.20 (2.21) 0.50 0.19 (0.19) 0.00 0.65 (1.1) 0.50

H13 (Cont.) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

H14 (TMX) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

H15 (CTD) 0.19 (0.19) 0.00 0.39 (0.58) 0.50 0.19 (0.19) 0.00

H16 (Cont.) 0.19 (0.19) 0.00 0.40 (0.61) 0.50 0.71 (1.23) 0.50

H17 (TMX) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

H18 (CTD) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

H19 (CTD) 0.19 (0.19) 0.00 0.51 (0.82) 0.50 0.19 (0.19) 0.00

H20 (Cont.) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

H21 (TMX) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

UK22 (TMX) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

UK23 (Cont.) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

UK24 (CTD) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

UK25 (Cont.) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

UK26 (TMX) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

UK27 (CTD) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

UK28 (CTD) 0.37 (0.55) 0.50 0.19 (0.19) 0.00 0.19 (0.19) 0.00

UK29 (Cont.) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

UK30 (TMX) 0.4 (0.6) 0.50 0.19 (0.19) 0.00 0.19 (0.19) 0.00

UK31 (Cont.) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

UK32 (TMX) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

UK33 (CTD) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

48

Table S12C: Nectar only Site (treat.) Clothianidin



Median (max.)

Prop. Samples > LOQ

Median (max.)

Prop. Samples > LOQ

Median (max.)

Prop. Samples > LOQ

G1 (Cont.) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

G2 (CTD) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

G3 (TMX) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

G4 (Cont.) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

G5 (CTD) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

G6 (TMX) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

G7 (Cont.) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

G8 (CTD) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

G9 (TMX) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00 H10* (Cont.) NA NA NA NA NA NA H11 (TMX) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

H12 (CTD) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

H13 (Cont.) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

H14 (TMX) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

H15 (CTD) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

H16 (Cont.) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

H17 (TMX) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

H18 (CTD) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

H19 (CTD) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

H20 (Cont.) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

H21 (TMX) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

UK22 (TMX) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

UK23 (Cont.) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

UK24 (CTD) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

UK25 (Cont.) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

UK26 (TMX) 0.73 (0.76) 1.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

UK27 (CTD) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

UK28 (CTD) 0.37 (0.54) 0.50 0.19 (0.19) 0.00 0.19 (0.19) 0.00

UK29 (Cont.) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

UK30 (TMX) 0.52 (0.84) 0.50 0.19 (0.19) 0.00 0.19 (0.19) 0.00

UK31 (Cont.) 0.86 (0.87) 1.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

UK32 (TMX) 0.19 (0.19) 0.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

UK33 (CTD) 0.78 (0.79) 1.00 0.19 (0.19) 0.00 0.19 (0.19) 0.00

*For H10 insufficient nectar was dissected from bees for a viable analysis of residues.

49

REFERENCES AND NOTES 1. A. J. Vanbergen, Insect Pollinators Initiative, Threats to an ecosystem service: Pressures on

pollinators. Front. Ecol. Environ 11, 251–259 (2013). doi:10.1890/120126

2. S. G. Potts, J. C. Biesmeijer, C. Kremen, P. Neumann, O. Schweiger, W. E. Kunin, Global pollinator declines: Trends, impacts and drivers. Trends Ecol. Evol. 25, 345–353 (2010). doi:10.1016/j.tree.2010.01.007 Medline

3. R. Winfree, R. Aguilar, D. P. Vázquez, G. LeBuhn, M. A. Aizen, A meta-analysis of bees’ responses to anthropogenic disturbance. Ecology 90, 2068–2076 (2009). doi:10.1890/08-1245.1 Medline

4. M. Henry, M. Béguin, F. Requier, O. Rollin, J.-F. Odoux, P. Aupinel, J. Aptel, S. Tchamitchian, A. Decourtye, A common pesticide decreases foraging success and survival in honey bees. Science 336, 348–350 (2012). doi:10.1126/science.1215039 Medline

5. P. R. Whitehorn, S. O’Connor, F. L. Wäckers, D. Goulson, Neonicotinoid pesticide reduces bumble bee colony growth and queen production. Science 336, 351–352 (2012). doi:10.1126/science.1215025 Medline

6. J. E. Cresswell, C. J. Page, M. B. Uygun, M. Holmbergh, Y. Li, J. G. Wheeler, I. Laycock, C. J. Pook, N. H. de Ibarra, N. Smirnoff, C. R. Tyler, Differential sensitivity of honey bees and bumble bees to a dietary insecticide (imidacloprid). Zoology 115, 365–371 (2012). doi:10.1016/j.zool.2012.05.003 Medline

7. B. A. Woodcock, N. J. B. Isaac, J. M. Bullock, D. B. Roy, D. G. Garthwaite, A. Crowe, R. F. Pywell, Impacts of neonicotinoid use on long-term population changes in wild bees in England. Nat. Commun. 7, 12459 (2016). doi:10.1038/ncomms12459 Medline

8. M. Rundlöf, G. K. S. Andersson, R. Bommarco, I. Fries, V. Hederström, L. Herbertsson, O. Jonsson, B. K. Klatt, T. R. Pedersen, J. Yourstone, H. G. Smith, Seed coating with a neonicotinoid insecticide negatively affects wild bees. Nature 521, 77–80 (2015). doi:10.1038/nature14420 Medline

9. G. E. Budge, D. Garthwaite, A. Crowe, N. D. Boatman, K. S. Delaplane, M. A. Brown, H. H. Thygesen, S. Pietravalle, Evidence for pollinator cost and farming benefits of neonicotinoid seed coatings on oilseed rape. Sci. Rep. 5, 12574 (2015). doi:10.1038/srep12574 Medline

10. D. Goulson, Neonicotinoids impact bumblebee colony fitness in the field; a reanalysis of the UK’s Food & Environment Research Agency 2012 experiment. PeerJ 3, e854 (2015). doi:10.7717/peerj.854 Medline

11. C. Sandrock, L. G. Tanadini, J. S. Pettis, J. C. Biesmeijer, S. G. Potts, P. Neumann, Sublethal neonicotinoid insecticide exposure reduces solitary bee reproductive success. Agric. For. Entomol. 16, 119–128 (2014). doi:10.1111/afe.12041

12. G. C. Cutler, C. D. Scott-Dupree, M. Sultan, A. D. McFarlane, L. Brewer, A large-scale field study examining effects of exposure to clothianidin seed-treated canola on honey bee colony health, development, and overwintering success. PeerJ 2, e652 (2014). doi:10.7717/peerj.652 Medline

50

http://dx.doi.org/10.1890/120126

http://dx.doi.org/10.1016/j.tree.2010.01.007

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=20188434&dopt=Abstract

http://dx.doi.org/10.1890/08-1245.1

http://dx.doi.org/10.1890/08-1245.1


http://dx.doi.org/10.1126/science.1215039


http://dx.doi.org/10.1126/science.1215025


http://dx.doi.org/10.1016/j.zool.2012.05.003


http://dx.doi.org/10.1038/ncomms12459


http://dx.doi.org/10.1038/nature14420


http://dx.doi.org/10.1038/srep12574


http://dx.doi.org/10.7717/peerj.854


http://dx.doi.org/10.1111/afe.12041

http://dx.doi.org/10.7717/peerj.652


13. G. Christopher Cutler, C. D. Scott-Dupree, A field study examining the effects of exposure to neonicotinoid seed-treated corn on commercial bumble bee colonies. Ecotoxicology 23, 1755–1763 (2014). doi:10.1007/s10646-014-1340-5 Medline

14. B. A. Woodcock, M. S. Heard, M. S. Jitlal, M. Rundlöf, J. M. Bullock, R. F. Shore, R. F. Pywell, Replication, effect sizes and identifying the biological impacts of pesticides on bees under field conditions. J. Appl. Ecol. 53, 1358–1362 (2016). doi:10.1111/1365-2664.12676

15. E. Pilling, P. Campbell, M. Coulson, N. Ruddle, I. Tornier, A four-year field program investigating long-term effects of repeated exposure of honey bee colonies to flowering crops treated with thiamethoxam. PLOS ONE 8, e77193 (2013). doi:10.1371/journal.pone.0077193 Medline

16. M. Henry et al., Reconciling laboratory and field assessments of neonicotinoid toxicity to honeybees. Proc. R. Soc. B Biol. Sci. 282, 2015.2110 (2015).

17. A. Jones, P. Harrington, G. Turnbull, Neonicotinoid concentrations in arable soils after seed treatment applications in preceding years. Pest Manag. Sci. 70, 1780–1784 (2014). doi:10.1002/ps.3836 Medline

18. C. Botías, A. David, J. Horwood, A. Abdul-Sada, E. Nicholls, E. Hill, D. Goulson, Neonicotinoid residues in wildflowers, a potential route of chronic exposure for bees. Environ. Sci. Technol. 49, 12731–12740 (2015). doi:10.1021/acs.est.5b03459 Medline

19. D. Goulson, An overview of the environmental risks posed by neonicotinoid insecticides. J. Appl. Ecol. 50, 977–987 (2013). doi:10.1111/1365-2664.12111

20. A. Fairbrother, J. Purdy, T. Anderson, R. Fell, Risks of neonicotinoid insecticides to honeybees. Environ. Toxicol. Chem. 33, 719–731 (2014). doi:10.1002/etc.2527 Medline

21. FERA, Neonicotinoid Pesticides and Bees. Report to Syngenta Ltd. (The Food and Environment Research Agency, UK, 2013).

22. F. Sánchez-Bayo, D. Goulson, F. Pennacchio, F. Nazzi, K. Goka, N. Desneux, Are bee diseases linked to pesticides? - A brief review. Environ. Int. 89-90, 7–11 (2016). doi:10.1016/j.envint.2016.01.009 Medline

23. C. R. Archer, C. W. W. Pirk, G. A. Wright, S. W. Nicolson, Nutrition affects survival in African honeybees exposed to interacting stressors. Funct. Ecol. 28, 913–923 (2014). doi:10.1111/1365-2435.12226

24. European Commission, Eurostat: Agriculture (EC, 2016); http://ec.europa.eu/eurostat/web/agriculture/data/database.

25. A. Imdorf, G. Buehlmann, L. Gerig, V. Kilchenmann, H. Wille, Überprüfung der Schätzmethode zur Ermittlung der Brutfläche und der Anzahl Arbeiterinnen in freifliegenden Bienenvölkern [A test of the method of estimation of brood areas and number of worker bees in free-flying colonies]. Apidologie (Celle) 18, 137 (1987). doi:10.1051/apido:19870204

26. D. Kleijn, R. Winfree, I. Bartomeus, L. G. Carvalheiro, M. Henry, R. Isaacs, A.-M. Klein, C. Kremen, L. K. M’Gonigle, R. Rader, T. H. Ricketts, N. M. Williams, N. Lee Adamson, J. S. Ascher, A. Báldi, P. Batáry, F. Benjamin, J. C. Biesmeijer, E. J. Blitzer, R. Bommarco,

51

http://dx.doi.org/10.1007/s10646-014-1340-5


http://dx.doi.org/10.1111/1365-2664.12676

http://dx.doi.org/10.1111/1365-2664.12676

http://dx.doi.org/10.1371/journal.pone.0077193


http://dx.doi.org/10.1002/ps.3836


http://dx.doi.org/10.1021/acs.est.5b03459


http://dx.doi.org/10.1111/1365-2664.12111

http://dx.doi.org/10.1002/etc.2527


http://dx.doi.org/10.1016/j.envint.2016.01.009


http://dx.doi.org/10.1111/1365-2435.12226

http://dx.doi.org/10.1051/apido:19870204

M. R. Brand, V. Bretagnolle, L. Button, D. P. Cariveau, R. Chifflet, J. F. Colville, B. N. Danforth, E. Elle, M. P. D. Garratt, F. Herzog, A. Holzschuh, B. G. Howlett, F. Jauker, S. Jha, E. Knop, K. M. Krewenka, V. Le Féon, Y. Mandelik, E. A. May, M. G. Park, G. Pisanty, M. Reemer, V. Riedinger, O. Rollin, M. Rundlöf, H. S. Sardiñas, J. Scheper, A. R. Sciligo, H. G. Smith, I. Steffan-Dewenter, R. Thorp, T. Tscharntke, J. Verhulst, B. F. Viana, B. E. Vaissière, R. Veldtman, K. L. Ward, C. Westphal, S. G. Potts, Delivery of crop pollination services is an insufficient argument for wild pollinator conservation. Nat. Commun. 6, 7414 (2015). doi:10.1038/ncomms8414 Medline

27. B. Magnusson, T. Näykki, H. Hovind, M. Krysell, Handbook for Calculation of Measurement Uncertainty (Nordtest report 537, Nordic Innovation, Norway, 2012).

28. T. Blacquière, G. Smagghe, C. A. van Gestel, V. Mommaerts, Neonicotinoids in bees: A review on concentrations, side-effects and risk assessment. Ecotoxicology 21, 973–992 (2012). doi:10.1007/s10646-012-0863-x Medline

29. European Food Safety Authority (EFSA), Conclusion on the peer review of the pesticide risk assessment for bees for the active substance clothianidin. EFSA J. 11, 3066 (2013). doi:10.2903/j.efsa.2013.3066

30. EFSA, Conclusion on the peer review of the pesticide risk assessment for bees for the active substance thiamethoxam. EFSA J. 11, 3067 (2013). doi:10.2903/j.efsa.2013.3067

31. EFSA, Conclusion on the peer review of the pesticide risk assessment for bees for the active substance imidacloprid. EFSA J. 11, 3068 (2013). doi:10.2903/j.efsa.2013.3068

32. R Core Development Team, R: Version 3.2.1 A Language and Environment for Statistical Computing (R Foundation for Statistical Computing, 2015); http://cran.r-project.org.

52

http://dx.doi.org/10.1038/ncomms8414


http://dx.doi.org/10.1007/s10646-012-0863-x


http://dx.doi.org/10.2903/j.efsa.2013.3066



Country-specific effects of neonicotinoid pesticides on honey bees and wild bees

J. Peyton, S. Hulmes, L. Hulmes, M. Sárospataki, C. Saure, M. Edwards, E. Genersch, S. Knäbe and R. F. PywellB. A. Woodcock, J. M. Bullock, R. F. Shore, M. S. Heard, M. G. Pereira, J. Redhead, L. Ridding, H. Dean, D. Sleep, P. Henrys,

DOI: 10.1126/science.aaa1190 (6345), 1393-1395.356Science

, this issue p. 1395, p. 1393; see also p. 1331Scienceaffect pollinator health under realistic agricultural conditions.and colony reproduction in both honeybees and wild bees. These field results confirm that neonicotinoids negativelyon rapeseed in Europe, find that neonicotinoid exposure from several nontarget sources reduces overwintering success

, in a multicounty experimentet al.especially when coexposed to a commonly used agrochemical fungicide. Woodcock exposed to neonicotinoids for 3 to 4 months via nontarget pollen, resulting in decreased survival and immune responses,

find that bees near corn crops areet al.neonicotinoids diminish bee health (see the Perspective by Kerr). Tsvetkov environmental conditions. Two studies, conducted on different crops and on two continents, now substantiate thatHowever, lingering criticism was that the studies did not represent field-realistic levels of the chemicals or prevailing

Early studies of the impacts of neonicotinoid insecticides on insect pollinators indicated considerable harm.Damage confirmed

ARTICLE TOOLS http://science.sciencemag.org/content/356/6345/1393

PERMISSIONS http://www.sciencemag.org/help/reprints-and-permissions

Terms of ServiceUse of this article is subject to the

is a registered trademark of AAAS.Sciencelicensee American Association for the Advancement of Science. No claim to original U.S. Government Works. The title Science, 1200 New York Avenue NW, Washington, DC 20005. 2017 © The Authors, some rights reserved; exclusive

(print ISSN 0036-8075; online ISSN 1095-9203) is published by the American Association for the Advancement ofScience

on June 29, 2017


Dow

nloaded from

http://science.sciencemag.org/content/356/6345/1393

http://www.sciencemag.org/help/reprints-and-permissions

http://www.sciencemag.org/about/terms-service


Documents

NEONICOTINOIDS Country-specific effects of neonicotinoid ...€¦ · the reproductive success of bumble bees (5, 8, 10) and solitary bees ( 8, 11); other studies have iden-tified